LCD8S and KT48SVPRK sine wave controller.

Ins0mniac

1 mW
Joined
Jun 29, 2022
Messages
11
Hey all,

I am deep in the guts of my controller and LCD8S display attempting a fix and thought it worth a shot to see if anyone here has anything similar and what they did to fix it. I have a new LCD8S display on the way as I think the issue is there, but it takes about a month from China so I figured I'd work on it in the meantime as summer is winding down and I might not get to ride as much in a month.

So here goes:

My controller and LCD8S worked just fine with a 48 ZWS KT controller (square wave) and still works fine with my old YunyanKing Controller. It's a cyclone 3kW motor with a 52V 14s7p battery plus now this 48 volt SVP controller.

So shortly after riding my ebike for a few loops around the area I live in (less than 6-7 min maxing it out) I put it inside and ended up with a fault in the controller. Turns out my 270 ohm resistor had burnt out and led to downstream failure of the LM317T that steps down voltage from what should have been 52v -> 35v or so across the resistor:

IMG_20220910_064215__01__01.jpg

Then from 35v in the LM317T input to 15V output which later goes on to the 5v stepdown section.

IMG_20220910_063327__01__01.jpg

Anyhow, I substituted this resistor for a buck converter stepdown module that worked for a bit:

IMG_20220913_093933.jpg

And I believe i ended up with a throttle short that somehow sent 56v to my throttle input red wire, perhaps related to a hall sensor fix that went faulty, i am thinking due to overheating the solder. This I fixed, replaced my throttle hall sensor, verified it works and speeds up the motor. At this point I attached my old YuyanKing controller and sped off zero problems at 35 mph down the street without voltage drops.

Now, I reattach my existing 48v SVP controller and even my 48v ZWS controller. I have replaced the downstream capacitor immediately after the resistor that is 270 ohm with one rated for higher voltage but same uf rating and replaced the resistor back and removed the buck converter module to eliminate possible errors and because I burnt two of them troubleshooting and well, resistors are cheaper.

I also replaced a few resistors on the LCD8S board that I determined had been slightly faulty after this short and I replaced the A1013 transistor on the LCD8S. I also wasn't sure about the status of a capacitor right above my Pin 1 above the A1013 and it read around 130 pf capacitance and I didn't have this size on hand, so replaced it with a 100 nf capacitor I had. The image below shows the resistor and capacitor I am speaking of outlined in a red box:

IMG_20220913_060222__01.jpg

This worked ok to turn on the LCD as without it, the LCD didn't stay on while pushing power. I had a 302 resistor above that capacitor that was burnt that I also replaced.

So onto the interesting part:

Now, my LCD8S works just fine, turns on, displays voltage and bars, and even throttle produces power output and gives me motor power. However, on this particular model of controller (I have yet to hard test my ZWS, but presume it will be the same result) the power is being cut after I put in about 100w or so of power. I have taken numerous throttle measurements and my input power is around 4.6 v and signal is about 0.8 to start and goes up as I push the throttle down. The issue is that this cutout happens about 3-5 seconds in. It seems to correlate with the display voltage reading about 3v less that prior to throttling. So I have traced the DM pin from the Transistor and it ends up connecting to Pin 18 or VSS_4 on the STM32 processor on the LCD. I have attached my diagrams.

IMG_20220915_064709__01__01__01__01__01.jpg

So it goes DM -> 104 resistor (100kOhm, reads spot on) -> 332 resistor (3.3kOhm - this resistor reads more like 3.15kOhm not sure if this is maybe contributing to my issues) -> Pin 18. I have circled the path and exit start from DM on the A1013 and ending on the STM32 pin.

I have done additional troubleshooting and have identified that DM on the Transistor reads at roughly the same voltage as base and emitter if the pin is cut. If soldered back to the pin so that it contacts the PCB it reads a lower voltage by about 3-4v.

Is it possible there is a capacitor connected to these pins that causes this drain to be greater than expected or that there is a short to a ground somewhere that eats around 3-4v?

So with this image below, the pin connections that have continuity with pin 18 (where DM traces to, through two resistors):

Screenshot_20220915-233932__01__01__01.jpg

And for ease of view I have included and labeled STM32 printout from the reference datasheet:

Screenshot_20220915-232940__01__01.jpg

Pin 18 on the processor is continuous with the following pins.

Start: Pin 18 (VSS_4) to:
Pin 12 (VSS_A)
Pin 63 (VSS_3)
Pin 60 (BOO T0)
Pin 47 (VSS_2)

This is shown on my PCB and in a diagram of the STM32 (linked above, for ease of viewing and troubleshooting):

Any suggestions on what could be causing this parasitic suck of voltage down 3 ish volts? The capacitors coming off the continuous pins actually do read around 3.3 volts on the positive side, the side not directly connected to the pins tho and not 3.3v on the side connecting to the processor, that is more of a low low voltage, I believe something like 0.08 volts or so after being stepped down from it's high of 56 volts or so throught the two resistors mentioned above.

Odd part is that the Transistor reads higher on the middle pin (56v or so) and this is my desired value. Once I connect the pin to the PCB though, it begins to read lower at 53 volts as if something downstream is sucking the voltage lower. This is where my issue begins I think. I could be wrong but it seems to be related to this.

I have also considered that maybe the high voltage on my throttle of over 4.2 (anywhere from 4.3-4.6) is causing issues and somehow preventing it to continue running. I have tried adjust various settings in the LCD programming menu also and this hasn't availed any different performance.

The motor makes a groan, and runs weakly for a bit then cuts out power. Again, with my other YunyanKing Controller it runs like a champ, given with a different phase combination. I have considered changing phase combos except that this combo worked just fine on my KT 48V ZWS until I had this resistor short on me. I was going 37mph plus and did around 200 miles before it failed on me. I do not see why switching to a sine wave would have done this. I have considered shorted Mosfets also but preliminary testing on board hasn't yielded obvious shorts, will now move to desoldering and testing each I guess.

If it helps, prior to doing anything, these were my voltage readings on the SVP controller stock from when I just received it, even before riding (voltages in red and yellow shown, switched halfway through as the yellow was easier to visualize):

IMG_20220910_064841__01__01.jpg

And the ADC converter that came stock from the factory and voltages of the pins as taken from the rear of the board:

IMG_20220909_124402.jpg

IMG_20220910_064423__01.jpg

I am stipulating that one of the following is occuring:

1) My ADC converter is faulty, introducing a phase issue, but this seems unlikely given there is still continuity on the phase without it via the resistor and the ADC bridge input, unless programming prevents it from functioning properly by activating another phase while this one is yet working.

2) Blown mosfet(s), will be testing one by one today.

3) Blown capacitor or voltage regulator on the LCD8S but I am not sure of what the reference values are, so this is difficult to troubleshoot. Despite my detailed diagrams and mild success, I am not an electrical engineer by trade 😆

4) Faulty throttle voltage as it's over 4.2v, unsure of real world implications of this spec, however and if that's what it's meant to be.

5) Slightly less functional resistor 332 between DM pin and VSS_4 on STM32 on LCD8S (reading 3.15 kOhm not 3.3 kOhm).

6) Faulty microprocessor either on controller, LCD8s or both.

7) ??????? Please suggest other things I may have overlooked.

This switch to sine wave has been somewhat of a headache. Please help.
 
https://endless-sphere.com/forums/viewtopic.php?f=2&t=114482&p=1692939&hilit=temperature#p1692939

Your controller is 36-48v? Well, then it was designed for 13S max. And by designed, I mean that's the absolute max under ideal conditions without being pushed. It's a typical no name Chinese product. Derate by 25-50%. :mrgreen:
 
Comrade said:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=114482&p=1692939&hilit=temperature#p1692939

Your controller is 36-48v? Well, then it was designed for 13S max. And by designed, I mean that's the absolute max under ideal conditions without being pushed. It's a typical no name Chinese product. Derate by 25-50%. :mrgreen:

I don't see why that would prevent it from working off the bat as 48v controllers are typically ran with 63v capacitors which is precisely what I have. In addition to the fact that my ZWS square wave is the same model exactly absent the ADC. The mosfets are rated for 85v and 150 amperes, which is well within my limits with a 50a bms current limit and 58.8 top voltage. Anyhow none of my capacitors on the board are testing bad, at least from what I've measured. I suspect the 10 uf 63v cap on the LCD8S but that part is still in route so won't be able to confirm until I take it off and replace it.

Any other suggestions? KT doesn't have 52v controllers so the only option for someone with a 52v battery is the 48v controller lol.
 
Ins0mniac said:
I don't see why that would prevent it from working off the bat as 48v controllers are typically ran with 63v capacitors which is precisely what I have.

Did you read my link? The resistor + LM317T is a linear regulator network. The higher the battery voltage, the more power that regulator network has to dissipate. It's not even robust enough for long term 48V in my estimation. Capacitors have nothing to do with this.
 
Comrade said:
Ins0mniac said:
I don't see why that would prevent it from working off the bat as 48v controllers are typically ran with 63v capacitors which is precisely what I have.

Did you read my link? The resistor + LM317T is a linear regulator network. The higher the battery voltage, the more power that regulator network has to dissipate. It's not even robust enough for long term 48V in my estimation. Capacitors have nothing to do with this.

Yes I read the whole thread. I think you missed the part of my post where I had replaced the LM317T entirely with a buck converter. I had the same issue. To be fair I think I might have included a different image where I had used the buck converter in front of the LM317T.

I'll clarify, I replaced the LM317T with a step down buck converter that dropped voltage from 58v to 15v and put that on the input and output pins of the previous spot of the LM317T. This still had the same issue.

In addition, checking the LM317T still yields stable 15V output. So it's not that.

To clarify: I haven't even rode 10ft with this controller. I simply air tested it and the motor stalls out a few seconds in with this sine wave drive controller. Previous square wave didn't have this issue, that was the one I drove for a few hundred miles before the resistor burnt up.

I am starting to suspect the amperages involved may not be programmed for triggering the phases properly. I found this mentioned by another user on another topic:

Arlo1 said:
Njay said:
Sorry, I was actually speaking from a controller development point of view :D

Mines always blow silently and suddenly, no smoke, no plasma, no bang-sound, no visible physical damage (I don't have current limiting active yet, but I know it's not always the current).
If a item is operated with in its parameters it should last for a very long time! SO if you have a good quality fet and you have it fail then one or more of the parameters must have been exceeded. If you run it at the max voltage hot off the charger then it will likely see voltage spikes above the max voltage rating. If you are using a china controller it will likely be using battery amps to guess at a phase amperage which is what the fet has to deal with. It is also possible the Caps are not located very well or they are china knock offs and are not doing the right job. The only way to find this is to test with the proper equipment and test in different operating scenarios until you find the parameter being exceeded. I
But when you say you don't have any exciting explosions like I do and they just "go to sleep" then I would bet they are just barely being pushed past one or more limits and might not be often. Or they are not a quality part or the inductive paths are slightly different. Is it the same fet always blowing???
 
So for future reference, a capacitor was at fault.

The LCD8S has a 10uf 63v surface mount capacitor that I replaced and it fixed my issue of voltage appearing to sag. I first had to change the A1013 transistor that had blown, and a resistor up on top of it as well.

I am still working on figuring out why there is a discrepancy of 2.7 volts less once pin DM is connected from the Transistor, but otherwise not present prior. If I keep Pin 1 and Pin 3(b) connected but don't connect pin DM, and read pin DM coming off the Transistor I get appropriate voltage. Once connecting, however, it drops by 2.7 volts less.

The other issue is with my sine wave controller, the 48 SVP version. It just doesn't work and there is now a short between ground and positive voltage after all my monkeying about with it. I shorted a mosfet between gate and drain while taking measurements (system had residual built up in capacitors) as I was trying to diagnose why there is only 54 Ω resistance between the ground pin of my ACS 780 and the 5V input pin.

On my square wave controller this resistance steadily rizes to a high value as I'm reading it with the multimeter, thinking that there is a capacitor in parallel with it. I have testing the capacitors I see with an in circuit ESR meter and hasn't yielded any that could be faulty, and visually they look good. In addition, it's unlikely to be a capacitor and more likely to be a resistor or mosfet.

I guess the only thing left is to desolder every mosfet and test one by one. The problem is now going to be finding a k150e09ne or tk150e09ne as they're sometimes labeled. Finding an equivalent mosfet also has proven to be difficult as going to alltransistors.com and putting in specs leads to either a different Rds value or lower heat dissipation, and even of those one or two models it gives as similar, those are also low in stock.

I have ordered another Sine wave controller with the hopes of this second one working better. I also broke the ACS module in an attempt to desolder it and diagnose if the low resistance between ground and 5v was coming from it's internals, so have an ACs 751 100B on the way that I hope will replace it just the same (for reference this 100B I have also installed in my square wave controller in anticipation of putting the custom firmware on it later and converting to a sine wave, but wanted these voltage issues ironed out before then and an actual sine wave controller to practice on, in case my ACS module wasn't up to spec for the translation of square to sine wave.

So three problems now:

1) shorted Mosfets, need new mosfets. I'd happily take suggestions on what ones would work with my 52v battery and 50 amp current limit on BMS. Do these Rds values all have to match some specific number for the processor to read it right or are they able to be swapped out to a different value as long as I make all of the mosfets the same?

2) voltage on LCD8S still reads 2.7 volts lower than actual battery voltage, but doesn't sag anymore. Any suggestions on fixing that?

3) problem with ground and 5v continuity on my non-functional controller being only 54 ohm while the functional controller had a steadily increasing resistance once measurement begun. Suggestions?

These controllers are virtually identical, down to the resistors everywhere. So you'd think it be easy to track down but I'm still grasping these circuits and slowly developing a good understanding of them.. maybe you or someone here has better experience with these things and can point me in the correct direction.

Thank you.
 
Ins0mniac said:
I am still working on figuring out why there is a discrepancy of 2.7 volts less once pin DM is connected from the Transistor, but otherwise not present prior. If I keep Pin 1 and Pin 3(b) connected but don't connect pin DM, and read pin DM coming off the Transistor I get appropriate voltage. Once connecting, however, it drops by 2.7 volts less.
I don't know the specific cause in your system, but generally:

If you're seeing a voltage drop of the battery voltage incoming to a system, only when the system is "on", then the system is drawing so much current that the supply can't maintain the voltage.

That can be from resistance between the supply and the point being measured (so the supply point voltage will still read correctly), or a problem with the supply itself (in which case the voltage will drop equally at the supply point as well as at the original measurement point)

It can be from a load within the system being too high for the supply, from a failed or failing component most commonly.

(supply being either the battery itself, or any regulation or switching system between the battery and the system, etc).

(system being whatever is being powered from the supply).





The other issue is with my sine wave controller, the 48 SVP version. It just doesn't work and there is now a short between ground and positive voltage after all my monkeying about with it. I shorted a mosfet between gate and drain while taking measurements (system had residual built up in capacitors) as I was trying to diagnose why there is only 54 Ω resistance between the ground pin of my ACS 780 and the 5V input pin.

Sounds like two separate problems.

The 5v low-resistance usually means one of the components powered by the 5v supply has been damaged internally. This can be the MCU itself, or any of the logic components, sensors, etc., powered by 5v. Sometimes it's the actual 5v regulator internally failed. Rarely, it's a leaky capacitor on the 5v supply.

Sometimes this problem is caused by a wiring fault that causes a 5v device or signal to be shorted to battery voltage, most commonly by a motor wiring fault, such as axle-end cable damage allowing phase wires to short to hall wires, or a throttle cable fault that allows a battery voltmeter or switch to short to throttle 5v supply or signal wires (or water internal to the throttle that does the same thing).



The battery voltage input to ground short is usually FETs, and can be any or all of them (usually just one phase). Replacing the FETs that failed is required; it's safer to replace all the FETs in that phase than just the ones that failed if they didn't all fail, to keep current sharing more equal (it's never perfect anyway), by using all the same FETs with the same RDSon.

If there's a huge difference in the Qg (gate charge) or other gate properties between the new FETs and any remaining ones being kept, it may also cause issues of timing turning on and off creating extra heat in the ones that turn on first, but it's not as big a deal as RDSon in most of these controllers.

Lower RDSon is better, less heat generated for the same current.

Higher Vds rating is better, more tolerance for voltage spikes in the system, but higher Vds often means higher RDSon, so it's a compromise.

If the new FETs are hugely different from the remaining ones, the controller may perform better if all of them are replaced.

It is often cheaper to buy a new controller than replace all the FETs with known-good non-counterfeit parts. (stuff you find on ebay, aliexpress, etc has a high risk of not being what it says it is; Mouser, Digikey, Farnell, etc are much more likely to have the real stuff).

Note the MCU in the controller doesn't read any FET values at all, and has no idea what's there--it just sends on/off signals to the gate drivers with timing derived from the hall sensor signals for sensored controllers, and derived from phase wires via other sensors for sensorless ones.

The battery current limit isn't directly relevant to the FET current--the FETs see motor phase current which can be mroe than 1.5x the battery current (sometimes much higher depending on controller design and system loading).


There are some threads around here with (partial) schematics of some controllers, most of them pretty old, but potentially useful for understanding how parts of them work, if you want to look for them (I don't have any links, but there may be some in the Sticky Index threads for each section).

There are also DIY-controller-design threads with more theory and design info that may be helpful, but they won't directly apply to the design specifics of what you're working with, for the most part.
 
Thank you for the information amberwolf!

I ended up installing this buck converter again in place of the resistor and something again backfired in my LCD8S. This makes me think that wiring a diode across the place of the former resistor might be worth looking into. I am unsure if my buck converter backed some current or capacitor charge into the LCD and fried something internally.

In any case, upon attempting to fix that, which required yet another A1013 Transistor change, I then tried to really get down to why my battery voltage was still reading lower than necessary and accidentally shorted some zener diodes on the LCD display board.

Upon further inspection, post transistor short, I found my VSS and VDD pins shorted on the STM32 controlling the LCD display. I diagnosed a faulty 3.3v converter at this point, that is part number CJT1117B for which I ordered part number AMS1117-3.3 to replace it with.

That wasn't the end of that though. As unsoldering the regulator led to the discovery of yet still shorted VSS and VDD pins which hadn't previously been shorted prior to the accidental diode short caused by measuring voltages.

I ended up testing everything and nothing seemed out of spec. So I ended up desoldering the STM32 microprocessor itself to see if it was a short within that.

Turns out it was. Shame too.

Do you or anyone here know if the STM32 is programmed? It seems like it has to be, as it must output numbers and letters onto the LCD display itself. But I was wondering if i replaced the microprocessor if this would just fix itself or if it's not worth the hassle.

So now I have a dead LCD display. But good news!! My LCD8S from China is arriving today 🥳

However, I am hesitant to use it yet, as I don't want these issues popping up in the new display.

I went ahead and installed an M4 rectifier diode rated for 400V 30A forward impulse current, which is overkill for this application, but it's what I had on hand from some old circuit boards I am cannibalizing and don't want to wait two days for a lower spec SMD diode.

So this is what it looks like now on my one functional controller (hopefully functional):

IMG_20220921_092917.jpg

I will update if this works once my parts arrive later this evening.

Thank you for the help. Still going to have to figure out what went wrong on the last controller board but may take a break for a bit as I am tired of diagnosing and breaking things and just want a functional bike at this point.
 
Good question. I am not sure.

I had assumed it just didn't withstand the test of time but I may be wrong. When this happened I had to change the LM317T and the resistor at the same time.

Then onto the A1013 in the LCD8S.

I would guess heat energy and my high voltage/current stressing the resistor. These things are just cheaply used to step the voltage down for the LM317T to be manageable.

When it happened I had just finished fixing my bikes chainrings and taken it for a top speed run around my place, for only about 150feet or so, and back.

So no hard riding for too long then. It turned off ok. I went to test it out again 30 minutes later and it wasn't turning on.
 
Ins0mniac said:
I ended up installing this buck converter again in place of the resistor and something again backfired in my LCD8S. This makes me think that wiring a diode across the place of the former resistor might be worth looking into. I am unsure if my buck converter backed some current or capacitor charge into the LCD and fried something internally.
I don't know if the diode will help, but it shouldn't hurt.

I expect it's more likely to be an overcurrent problem:

If you're using the display's power-on function to turn the controller on and off, keep in mind that it does this by passing the battery power that comes to it from the controller back on a separate wire to the controller's "KSI" keyswitch / ignition line, via a small transistor controlled by the display's MCU. That KSI line is where battery power comes from to run the buck converter (or the big resistor) to power the LVPS system that runs the brain and accessories and gate drive/etc of the controller. (battery power to the FETs is never turned off at all except by your battery or any other main power switch or disconnect you may have).

It's possible the buck converter (we'll call it a DC-DC for short) is drawing more current thru the LCD's switch electronics (usually a very tiny transistor or FET), and damaging that.

Or if it uses switching (SMPS) technology, and there is insufficient filtering (capacitors) on input and output, it might be making so much electrical noise that it causes problems with the display's own LVPS (which may also be an SMPS with little or no filtering), leading to failures.




Do you or anyone here know if the STM32 is programmed? It seems like it has to be, as it must output numbers and letters onto the LCD display itself. But I was wondering if i replaced the microprocessor if this would just fix itself or if it's not worth the hassle.

Yes, it's programmed (at the display factory) to do everything that it does (the STM32 itself is just a single-chip computer, or MCU, and doesn't do anything until you write software for it), but there's no way I know of for you to reprogram it with the same firmware, as the FW itself isn't available that I know of, for any of these. (and it's likely they have set protection up against reading an already-programmed one to get it from there).

Some displays have open-source firmware replacements (OSFW), some written here on ES in various threads, but I think to use those you'd also have to use the matching OSFW on the controller (if there is one for yours). You'd have to read the thread(s) to find out, as I haven't followed them in any detail. :(


But I'd say that with the various chains of problems you're having, there is probably damage to components that haven't failed yet, but will and may cause further problems with other things (including new parts) before they do fail (if they fail). So you're probably better off starting anew. ;)

I've had "cursed" devices before, where they just can't be fixed and keep going wrong in different ways over and over...sometimes it just sucks and I've had to start over with a new device (of whatever kind, simple or complex, hardly any of them ebike-related). It's often power-supply (or power-filtering) related, but not always.

Thank you for the help. Still going to have to figure out what went wrong on the last controller board but may take a break for a bit as I am tired of diagnosing and breaking things and just want a functional bike at this point.
Remember: If it's not broken, improve it till it is. ;)
 
Ins0mniac said:
Good question. I am not sure.

I had assumed it just didn't withstand the test of time but I may be wrong. When this happened I had to change the LM317T and the resistor at the same time.
If you're using a voltage higher than the controller was built for (regardless of what the seller *says* it can do), the resistor will have higher voltage across it, which even if it's the same current thru it it must dissipate more power as heat, and the LM317 could have a higher voltage at it's input. At some point the heat generated in one or the other can cause it to fail, and often enough take the other one out by causing even higher heat generation in it.

Sometimes they simply don't use a big enough resistor (high enough wattage) for the power dissipation it must handle, especially within the controller's already-hot interior), so it may gradually degrade over time (faster the higher the system voltage).


You can calculate a new resistor value to replace the original one, if you want to stick with the big-dropping-resistor, so that it keeps the correct voltage at the input to the LM317 without overheating the resistor, for whatever your pack's full voltage is.

Alternately, using a DC-DC with sufficient filtering on it's input and output (if it's SMPS type), and of sufficient current- and power-handling ability, would work. If it's a linear type, it doesn't need filtering as it doesn't create it's own noise (but it may create more waste heat and that might be a problem inside the controller where it can get toasty already).

Another thought: There have been problems with counterfeit chips on some of those small DC-DC boards, where they have the p/n for a chip that can handle the voltage /etc but they aren't really that chip (or are defective or "seconds" in some way). The problem seen most often is the input voltage can't go as high as it should be able to (and it will eventually fail if used beyond whatever it's actual limit is, which of course can only be determined by turning it up till it dies :lol: ).


FWIW, this higher current must also then be handled by the display's transistor that switches power to the controller when the dipslay's power switch is used.

Also, if you're using higher voltage than the display was built for (again, regardless of what the seller says), then the regulator electronics in the display (it's LVPS) are stressed, and may fail.


Too-often, linear types of LVPS can fail in a way that passes at least part of the original higher voltage thru to the devices powered by the LVPS, and most of those devices simply can't handle even a little bit more voltage than they were already getting.

It may not even fail, but simply be unable to regulate down to the level it was supposed to, and let slightly higher voltage thru than it should, if the higher voltage is an edge case near it's limits or just beyond them, rather than far enough beyond to just blow it up.

If the display's LVPS is linear (like the controller's is), and any of those things happened, it could explain some of the other damage you've seen.
 
amberwolf said:
I expect it's more likely to be an overcurrent problem:

If you're using the display's power-on function to turn the controller on and off, keep in mind that it does this by passing the battery power that comes to it from the controller back on a separate wire to the controller's "KSI" keyswitch / ignition line, via a small transistor controlled by the display's MCU. That KSI line is where battery power comes from to run the buck converter (or the big resistor) to power the LVPS system that runs the brain and accessories and gate drive/etc of the controller. (battery power to the FETs is never turned off at all except by your battery or any other main power switch or disconnect you may have).

It's possible the buck converter (we'll call it a DC-DC for short) is drawing more current thru the LCD's switch electronics (usually a very tiny transistor or FET), and damaging that.

Or if it uses switching (SMPS) technology, and there is insufficient filtering (capacitors) on input and output, it might be making so much electrical noise that it causes problems with the display's own LVPS (which may also be an SMPS with little or no filtering), leading to failures.

So this is the buck converter I am using:
Screenshot_20220922-072245__01.jpg

And it's description states:
Screenshot_20220922-072253__01.jpg

Found on Amazon:
https://amzn.to/3f9VQWK

This is as good of a pic as I can get of it, should hopefully be pretty clear:

IMG_20220922_073653.jpg

So am I correct to assume this is a linear converter? It doesn't appear to have as much heat loss as my resistor. It states it uses mosfets. I found a NCE0140KA mosfet and Baoter 3296 to step the voltage down. Something about the converter doesn't make sense to me. The mosfet Vgs says it's +/-20v and if I'm stepping down from 58v to 15v as I am, isn't that a massive 43v Vgs and outside the spec of the mosfet?

I've had a few of these fail and I had previously assumed it was from my LCD8S failing internally but I am starting to suspect this Vgs discrepancy now.

Do you have a good suggestion of an SMTP converter I can buy? Otherwise, per all this it seems safer to just replace the resistor. I bought a pack of those also, so no issues doing that. The only qualm is the heat loss plus impending failure of the LM317T but I suppose those are cheaper to replace than a LCD8S everytime 😆
 
So this morning I tore apart my brand new LCD8S and measured all the voltages I could find, minus the USB charger module which is separately attached to the rear of the display casing.

Here it is.

IMG_20220922_090355__01__01__01__01__01__01__01.jpg

I hope it services someone who stumbles upon this sometime. Personally, Id suggest just buying a new one now 😆 but if it helps, it helps. The time factor between getting one from China and having one in hand is major too.
 
Comrade said:
amberwolf said:
Or if it uses switching (SMPS) technology, and there is insufficient filtering (capacitors) on input

Or maybe it's very sufficient and there is large inrush current at initial startup. :mrgreen:
Yeah, or that. :) :bolt: :flame: But I suspect it doesn't have any at all, and depends on external components for that (which sellers of it don't know about, don't care about, and don't tell you about).
 
Ins0mniac said:
So this is the buck converter I am using:
<snip>
So am I correct to assume this is a linear converter?
I'm pretty sure that's an SMPS type (switched mode power supply). In general they work by a bit like a motor controller, but using FET to switch the incoming voltage and feed that into a transformer (instead of a motor), and the output side of the transformer is wired to give higher (boost) or lower (buck) voltage, then the output is filtered (capacitors, inductors) and sometimes further regulated; there's usually a feedback loop from output to a control chip to monitor output and keep it stable (at whatever it is set to, if adjustable) under load.

It doesn't appear to have as much heat loss as my resistor. It states it uses mosfets. I found a NCE0140KA mosfet and Baoter 3296 to step the voltage down. Something about the converter doesn't make sense to me. The mosfet Vgs says it's +/-20v and if I'm stepping down from 58v to 15v as I am, isn't that a massive 43v Vgs and outside the spec of the mosfet?
It depends on how the circuit is designed; there are a number of ways to make a switching supply. I don't know how they made theirs.

If they *are* using an insufficient FET then it can cause it to operate incorrectly or fail, potentially passing more voltage thru than it should depending on operating/failure mode (which would then blow up things on the output side that can't handle that).

FWIW, the LM2696
https://www.ti.com/lit/gpn/lm2696
and it's relatives like the 2596
https://www.ti.com/product/LM2596
are commonly-counterfeited parts in buck converters, and the bad ones can't take the input voltage the real ones can (only 24v for that version) (and/or the output current (only 3A)). Even the ones that are real chips still only handle a pretty low input voltage, so converters using them for higher voltages have to do something to bring it down to less than what they can handle, just like the LM317.

I can't remember the p/n I've seen most commonly, but it's in the same family. I'll edit it in here if I can find it.


Do you have a good suggestion of an SMTP converter I can buy?
Not specific ones. There are a number of companies that make good DC-DC units, that can be found on Mouser, Digikey, Farnell, etc. The ones I've used are all sealed potted PCB-mount units like Vicor, etc. (like this C&D one I'm using for 12v in my Lebowki-Honda controller hack:
https://endless-sphere.com/forums/viewtopic.php?t=105711
20200425_121421[1].jpg )



They generally look like the above, or these:
https://www.google.com/search?q=vicor+dc-dc&tbm=isch
but there are a lot of variations depending on power level required, and some are open-frame (no enclosure), or just plastic-potted, etc.

This is a Mouser search that shows a general 2-output device that takes nominally 48v (some of them a much wider range) and provides both a 5v line and a 12v line
https://www.mouser.com/c/power/dc-dc-converters/isolated-dc-dc-converters/?input%20voltage%2C%20nominal=48%20V&number%20of%20outputs=2%20Output&output%20voltage-channel%201=5%20V&output%20voltage-channel%202=12%20V

if you need other voltages instead (like 5v and 15v, depends on the gate driver supply voltage), you can specify those in their search, or you can use their chat function to ask for help finding the part you need.

You can get single-output ones instead, that do what you're already doing, just dropping the battery voltage to something the existing regulator system can use.

This mouser search uses a similar input range and outputs just 24v for the LM317.
https://www.mouser.com/c/power/dc-dc-converters/isolated-dc-dc-converters/?input%20voltage%2C%20nominal=48%20V&number%20of%20outputs=1%20Output&output%20voltage-channel%201=24%20V
 
Wow these are much smaller and way better of a form factor for this need.

Thanks Amberwolf!!

So now I have another question, is the 15V supply necessary for other components on the KT controller or is it's sole purpose to be stepped down to 5V?

From what I had gather by measuring voltages and looking at old similar styled controller schematics I understood it to be for stepping down the voltage to 5V via a cheaper transistor. Am I understanding this correctly?

Just trying to get the minimal output dc-dc converter from mouser now, and not something with more pins than I'll need to use. I guess I can always input it prior to the 5V converter but if I can elimate extra potentially failing components I'd wanna do that.
 
Ins0mniac said:
So now I have another question, is the 15V supply necessary for other components on the KT controller or is it's sole purpose to be stepped down to 5V?
probably the gate drivers, possibly other things.
 
Back
Top