steveo
100 kW
ZapPat said:
Does anyone know where i can purchase?.. I've tried searching the web, however non of the major distriutors carry this part!
-steveo
24 Mosfet Controller XJC8b116 microchip Information Thread
- Thread starter steveo
- Start date
Does anyone know where i can purchase?.. I've tried searching the web, however non of the major distriutors carry this part!
-steveo
steveo
100 kW
Oh,
and i'm looking for 50 or 24pieces .. not a big order, i will be testing the new mosfet!
-steveo
and i'm looking for 50 or 24pieces .. not a big order, i will be testing the new mosfet!
-steveo
methods
1 GW
When you are ready to run with the big dogs I can get anything from IR at the lowest prices anywhere.
Minimum quantity of 100. You wont start to see big savings till you get up over 300.
Sell off what you dont need on the board.
I would stick with IR Parts - even if another part looks slightly better on the DataSheet IR is well known in the industry for under-rating.
IR markets to Military and Aerospace where a rigorous qualification process is required - IR parts are preferred since we know they always exceed their ratings.
This is part of why we can run the IRFB4110 at 100V - I would not try that with other "100V" mosfets.
-methods
Minimum quantity of 100. You wont start to see big savings till you get up over 300.
Sell off what you dont need on the board.
I would stick with IR Parts - even if another part looks slightly better on the DataSheet IR is well known in the industry for under-rating.
IR markets to Military and Aerospace where a rigorous qualification process is required - IR parts are preferred since we know they always exceed their ratings.
This is part of why we can run the IRFB4110 at 100V - I would not try that with other "100V" mosfets.
-methods
I don't know all of what features the other controller has, but there are a couple of things I'd like to have in something like this:steveo said:
--variable regen, activated by:
1) squeezing the brake lever and then using the throttle to control the amount, assuming just standard switch-style brake handles (as I have now).
2) Additional method is having a second hall throttle type input that goes to the brake control, for those with sensors mounted on the brake handle for linear control directly instead of switch/throttle.
But it would need both, so it could be used with either type of bike. (I have only the first type of brake handle now, but later will build the second).
--Watt limiting. Not just current limiting, but actively tracking the average voltage the motor is at and limiting the current based on that voltage so it doesn't exceed a certain wattage. That would be a lot more effective in limiting the total power output of a motor to keep from destroying it, and also in being "legal" in some jurisdictions, than just a current limit.
Then why not see if they will sell you a sample of them directly for your new business venture?steveo said:
BTW, Digikey lists them
http://parts.digikey.com/1/parts/index22377.html
but has none in stock. You'd have to ask them if you haven't already.
Same with Mouser:
http://www.mouser.com/ProductDetail/Infineon/IPI075N15N3-G/?qs=rr3AYcQpCdeNmIvGU92nNQ%3d%3d
Knuckles
10 kW
Looks like XieChang may also have a 36-Fet controller available soon.
Poking around their docs I found a new "Design Parameter" flash application that shows the new EB224 board as well as a new EB236!
http://98.131.176.65/endless-sphere/V090914-214-100A.exe
COOL!
Poking around their docs I found a new "Design Parameter" flash application that shows the new EB224 board as well as a new EB236!
http://98.131.176.65/endless-sphere/V090914-214-100A.exe
COOL!
steveo
100 kW
Knuckles said:Looks like XieChang may also have a 36-Fet controller available soon.
Poking around their docs I found a new "Design Parameter" flash application that shows the new EB224 board as well as a new EB236!
http://98.131.176.65/endless-sphere/V090914-214-100A.exe
COOL!
36 mosfet
I think we are getting overboard on the mosfets now LOL.. 24mosfet with irfb4110 installed is easily good for 200amps contious with proper cooling ..doc did 100amps continous on a 12 mosfet crystalite ... !!!
-steveo
Knuckles
10 kW
Hey Steveo,
Do you speak "XieChang-ese"?
What's the deal-e-o with the new boards?
Inquiring minds want to know! :wink:
btw ... I want 120V 300A minimum as I have found a new friend ...
(ineeda cheap cheap mega-controller 4 my new friend
)
Do you speak "XieChang-ese"?
What's the deal-e-o with the new boards?
Inquiring minds want to know! :wink:
btw ... I want 120V 300A minimum as I have found a new friend ...
(ineeda cheap cheap mega-controller 4 my new friend
) Knuckles
10 kW
High res version here ... http://98.131.176.65/endless-sphere/24-fet_pcb.jpg
View attachment 2High res version here ...
http://98.131.176.65/endless-sphere/24-fet_100v.jpg
This is a prototype controller ...
24 - irfb4130 mosfets and
6 - 100v 1000uf main caps
phase and battery wires will be 10 AWG
btw ...
I don't think the power supply will dissipate the heat you mentioned.steveo said:
I suspect the power supply (voltage regulator) uses some kind of switching scheme.
Still trying to figure this out ...
But the shunts may get VERY HOT at 100 amps!
Each “M†shaped shunt is about 0.0076 ohm per unit.
(reference … http://endless-sphere.com/forums/viewtopic.php?f=16&t=7361)
1 shunt = 0.0076 ohm
2 shunts = 0.0038 ohm
3 shunts = 0.0025 ohm
4 shunts = 0.0019 ohm
For 100 amps with a shunt = .0019 ohm…
Power = I^2xR = 100x100x0.0019 = 19 watts YIKES!
(This is why methods modded the shunts on his 100 amp 18-fet model).
Doctorbass
100 GW
Thanks for these pics Knuckles,
I can't find any info on the IRFB4130.. Have you acces to the datasheet?
Would you better talk about the IRF4110?
Doc
I can't find any info on the IRFB4130.. Have you acces to the datasheet?
Would you better talk about the IRF4110?
Doc
Knuckles
10 kW
http://98.131.176.65/endless-sphere/irfb4110pbf.pdf
http://98.131.176.65/endless-sphere/irfb4310.pdf
http://98.131.176.65/endless-sphere/irfb4115pbf.pdf
Keywin only stocks the irfb4310 as he is so concerned about counterfeit mosfets in China (go figure).
But heck ... very easy to buy them here in North America and send them to Keywin for installation if you want ...
Considering REAL IR mosfets are ONLY manufactured in California and Mexico it kind of makes sense.
These are mosfet prices (per 1000) from IR's web site ...
(I think methods can get even cheaper than this for smaller quantities I am sure)
http://98.131.176.65/endless-sphere/irfb4310.pdf
http://98.131.176.65/endless-sphere/irfb4115pbf.pdf
Keywin only stocks the irfb4310 as he is so concerned about counterfeit mosfets in China (go figure).
But heck ... very easy to buy them here in North America and send them to Keywin for installation if you want ...
Considering REAL IR mosfets are ONLY manufactured in California and Mexico it kind of makes sense.
These are mosfet prices (per 1000) from IR's web site ...
(I think methods can get even cheaper than this for smaller quantities I am sure)
liveforphysics
100 TW
Those prices on IR's selling-to-the-public thing are not so good...
It seems this is the best option for the 150V folks:
IXTV130N15T
They are $1.4x in lots of 2,500 but they were not willing to break a lot. That's a cheap FET. Kinda a high volume commitment though. You would need to be looking to move ~100 of these 24 FET controllers in 150V trim before it would make sense.
Also, I don't trust IXYS specs in the same way I trust IR's specs. IR will quote an Rdson of 4mOhm on something, and it's the absolute max you're going to see if you bought 10,000 of them. Most of them are going to be like 2.5-3mOhm (just from my observation). If the IR part says 150v, I'm not to worried going a little above, if an IXYS part says 150v, I will be trying to keep it <~140v.
It seems this is the best option for the 150V folks:
IXTV130N15T
They are $1.4x in lots of 2,500 but they were not willing to break a lot. That's a cheap FET. Kinda a high volume commitment though. You would need to be looking to move ~100 of these 24 FET controllers in 150V trim before it would make sense.
Also, I don't trust IXYS specs in the same way I trust IR's specs. IR will quote an Rdson of 4mOhm on something, and it's the absolute max you're going to see if you bought 10,000 of them. Most of them are going to be like 2.5-3mOhm (just from my observation). If the IR part says 150v, I'm not to worried going a little above, if an IXYS part says 150v, I will be trying to keep it <~140v.
steveo
100 kW
Hey Everyone,
ok i was worried i must admit .. the forum being down for 2days ... lol
anyways
i have a new pic showing the 200v caps installed ..
i know they are all 470uf caps instead of 1000uf...
however .. i have no doubt in my mind that even though the uf is gona be lower with the 150v version of this controller.. as apposed to the higher uf on the caps at 100v .. they will be just a reliable!!
My ghetto controller i build used only 3 1000uf caps rated at 200v!! .. and it never had any issues!!
off the record .. even if you check crystalite old school box controller .. it has less then 1000uf total on all the caps!!
My 2 test controllers will be sent out to me monday!! .. i can't wait!
-steveo
ok i was worried i must admit .. the forum being down for 2days ... lol
anyways
i have a new pic showing the 200v caps installed ..
i know they are all 470uf caps instead of 1000uf...
however .. i have no doubt in my mind that even though the uf is gona be lower with the 150v version of this controller.. as apposed to the higher uf on the caps at 100v .. they will be just a reliable!!
My ghetto controller i build used only 3 1000uf caps rated at 200v!! .. and it never had any issues!!
off the record .. even if you check crystalite old school box controller .. it has less then 1000uf total on all the caps!!
My 2 test controllers will be sent out to me monday!! .. i can't wait!
-steveo
Knuckles
10 kW
Steveo,
Those are GREAT 200v caps. The best I could find in 18x40 size was 330uf.
Those are definitely 200v 470uf so you got GOOD stuff there.
I also like the way the caps are located on the EB224 board. Major design improvement IMO.
The power traces on the bottom of the board are nice and big too!
My concerns are ...
1) The voltage regulator circuit above 100 volts (note the 100v diode? "D8"="SS110")
2) Excessive shunt heating above 100 amps (add lots o solder maybe?)
3) Phase and hall wires at 8 AWG would be nice (will 8 AWG fit?)
4) Software settings may need to be "tricked" LVC, max amps, phase amps, etc.
5) Still need to tweak the regen HVC (still R12 maybe?) for regen above 75v
Those are GREAT 200v caps. The best I could find in 18x40 size was 330uf.
Those are definitely 200v 470uf so you got GOOD stuff there.
I also like the way the caps are located on the EB224 board. Major design improvement IMO.
The power traces on the bottom of the board are nice and big too!
My concerns are ...
1) The voltage regulator circuit above 100 volts (note the 100v diode? "D8"="SS110")
2) Excessive shunt heating above 100 amps (add lots o solder maybe?)
3) Phase and hall wires at 8 AWG would be nice (will 8 AWG fit?)
4) Software settings may need to be "tricked" LVC, max amps, phase amps, etc.
5) Still need to tweak the regen HVC (still R12 maybe?) for regen above 75v
steveo
100 kW
-Thanks for pointing out that 100v diode, i will look into this.. I've been told all we have to do is change out that big resistor i believe to 500ohms to accomidate 150v use
-i would not worry about the shunt, i've build my own diy controller before, and i've run 133v @ 50 amps .. and i never melted the shunt, infact, that controller is still in use
-i did discuss using 8 gauge instead of 10 gauge, however the factory 10 gauge wire is better quality, and can handle the heat if the wires where to run a bit warm.
-hvc/lvc/regen, i'm assuming modification is nessasary to accomidate this!!!
-steveo
-i would not worry about the shunt, i've build my own diy controller before, and i've run 133v @ 50 amps .. and i never melted the shunt, infact, that controller is still in use
-i did discuss using 8 gauge instead of 10 gauge, however the factory 10 gauge wire is better quality, and can handle the heat if the wires where to run a bit warm.
-hvc/lvc/regen, i'm assuming modification is nessasary to accomidate this!!!
-steveo
Knuckles said:Steveo,
Those are GREAT 200v caps. The best I could find in 18x40 size was 330uf.
Those are definitely 200v 470uf so you got GOOD stuff there.
I also like the way the caps are located on the EB224 board. Major design improvement IMO.
The power traces on the bottom of the board are nice and big too!
My concerns are ...
1) The voltage regulator circuit above 100 volts (note the 100v diode? "D8"="SS110")
2) Excessive shunt heating above 100 amps (add lots o solder maybe?)
3) Phase and hall wires at 8 AWG would be nice (will 8 AWG fit?)
4) Software settings may need to be "tricked" LVC, max amps, phase amps, etc.
5) Still need to tweak the regen HVC (still R12 maybe?) for regen above 75v
Since the server reset lost a lot of replies to this thread, here's mine in case it had been useful (I know there was a reply to it but can't even remember if I was able to finish reading it!):
If I were designing this, I'd lose the shunts entirely, and go with hall sensors reacting to the current in the power traces or something.
Or sense the voltage across the MOSFETs and calculate current based on the IR drop that results from, doing whatever math is needed to figure out what it is total (since that would be per-phase-pair), etc.
Of course, that only measures motor current, so if you're trying to measure battery current you'd need to go with the first method.
I wonder what a thermal imaging video of this board in operation would look like, especially at low speeds with high loads?
If I were designing this, I'd lose the shunts entirely, and go with hall sensors reacting to the current in the power traces or something.
Or sense the voltage across the MOSFETs and calculate current based on the IR drop that results from, doing whatever math is needed to figure out what it is total (since that would be per-phase-pair), etc.
Of course, that only measures motor current, so if you're trying to measure battery current you'd need to go with the first method.
I wonder what a thermal imaging video of this board in operation would look like, especially at low speeds with high loads?
methods
1 GW
steveo said:
And considering that they pop like popcorn in the microwave I would not consider that a good benchmark to measure my design by.
The 18 fet controllers from xlyte have 4000uF running in them - that is what you should consider as a starting point.
The IRFB4115GPBF in small quantities are $2.50 + tax + shipping from IRTronix btw.
-methods
steveo
100 kW
Hey guys,
Update....
-My 2 kits should be here by end of this week!
-I've ordered a sample batch of 48 mosfets for my 2 controllers!
should be hear by end of week
-steveo
Update....
-My 2 kits should be here by end of this week!
-I've ordered a sample batch of 48 mosfets for my 2 controllers!
should be hear by end of week
-steveo
liveforphysics
100 TW
Keep in mind, the quantity of capacitance is irrelevant to the function of the function and performance of the controller. In some ways it's more of a drawback than an advantage. What matters is the ESR of the caps. 100uF caps with 10mOhm ESR would perform better than 10,000uF caps with 11mOhm ESR. Don't get hung-up on the wrong value to look for when choosing a cap.
steveo
100 kW
liveforphysics said:Keep in mind, the quantity of capacitance is irrelevant to the function of the function and performance of the controller. In some ways it's more of a drawback than an advantage. What matters is the ESR of the caps. 100uF caps with 10mOhm ESR would perform better than 10,000uF caps with 11mOhm ESR. Don't get hung-up on the wrong value to look for when choosing a cap.
how would i find the esr rating of those china capasitors lol .. the ones in my 18 mosfet infinion are doing great!!
-steveo
Doctorbass
100 GW
Steveo, I agree with LFP,
For the ESR, it will protect your mosfetsand give you better performances and less heat!.. And the Capacitors will last longer
for my 18 fets controller for the speed record, i have bought from digikey some low 40mm x 22mm of 160V and THEY DO THE JOB !! Since many of us used 100V uf caps with 100V setup.. orsometime 105V... and never reported problems, I think that 160V caps with 150V max setup should be ok.
I measured the ESR of the Panasonic caps i bought and they are very low esr! I used a 12000$ Agilent analyst device dor IR and ESR. I observed that the stock crystalyte caps was pretty good ESR but not as good as the one i ordered.
i'll try to find the thread where i talked about them.
Edit:
The thread seem no longer availlableso i found it on the digikey website:
1000uF 160V 2000hrs and 105 degree ( -40 +105C)
it's the TS-HC serie..I remember i measured a ESR of 1/3 of the best crystalyte or Infineon stock caps value !!
they are 40mm x 22mm and over the qty i ordered the capacity and ESR was pretty stable.
Great Quality Panasonic caps
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=P13950-ND
Doc
For the ESR, it will protect your mosfetsand give you better performances and less heat!.. And the Capacitors will last longer
for my 18 fets controller for the speed record, i have bought from digikey some low 40mm x 22mm of 160V and THEY DO THE JOB !! Since many of us used 100V uf caps with 100V setup.. orsometime 105V... and never reported problems, I think that 160V caps with 150V max setup should be ok.
I measured the ESR of the Panasonic caps i bought and they are very low esr! I used a 12000$ Agilent analyst device dor IR and ESR. I observed that the stock crystalyte caps was pretty good ESR but not as good as the one i ordered.
i'll try to find the thread where i talked about them.
Edit:
The thread seem no longer availlableso i found it on the digikey website:
1000uF 160V 2000hrs and 105 degree ( -40 +105C)
it's the TS-HC serie..I remember i measured a ESR of 1/3 of the best crystalyte or Infineon stock caps value !!
they are 40mm x 22mm and over the qty i ordered the capacity and ESR was pretty stable.
Great Quality Panasonic caps
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=P13950-ND
Doc
Knuckles
10 kW
GENERAL PURPOSE (cheapo)
ALUMINUM ELECTROLYTIC CAPACITORS
Standard Case Profile
6, 12 and 18 mosfet XieChang controllers (Max Vertical Caps 16X25)
· Nichicon (VK) 100V-470uf 16X25 (1000ma Ripple - 85°C 1000 hours)
· Rubycon (YK) 100V-470uf 16X25 (950ma Ripple - 85°C 2000 hours)
· Nichicon (VZ) 100V-470uf 16X25 (715ma Ripple - 105°C 1000 hours)
· Nichicon (VY) 100V-470uf 16X25 (715ma Ripple - 105°C 1000 hours)
· Nichicon (VK) 200V-100uf 16X25 (450ma Ripple - 85°C 1000 hours)
· Rubycon (YK) 200V-100uf 16X25 (400ma Ripple - 85°C 2000 hours)
High Case Profile
24 and 36 mosfet XieChang controllers (Max Vertical Caps 18X40)
· Rubycon (YK) 100V-1000uf 18X40 (1350ma Ripple - 85°C 2000 hours)
· Nichicon (VY) 160V-470uf 18X40 (855ma Ripple - 105°C 1000 hours)
· Rubycon (YK) 160V-330uf 18X35.5 (760ma Ripple - 85°C 2000 hours)
· Rubycon (YK) 200V-330uf 18X35.5 (680ma Ripple - 85°C 2000 hours)
· Nichicon (VY) 200V-330uf 18X40 (675ma Ripple - 105°C 1000 hours)
The reported ripple current relates to the ESR ... discussion welcome.
Unfortunately 22x40 caps really don't fit on the 24-mosfet board ... or maybe they will ... Hmmm.
ALUMINUM ELECTROLYTIC CAPACITORS
Standard Case Profile
6, 12 and 18 mosfet XieChang controllers (Max Vertical Caps 16X25)
· Nichicon (VK) 100V-470uf 16X25 (1000ma Ripple - 85°C 1000 hours)
· Rubycon (YK) 100V-470uf 16X25 (950ma Ripple - 85°C 2000 hours)
· Nichicon (VZ) 100V-470uf 16X25 (715ma Ripple - 105°C 1000 hours)
· Nichicon (VY) 100V-470uf 16X25 (715ma Ripple - 105°C 1000 hours)
· Nichicon (VK) 200V-100uf 16X25 (450ma Ripple - 85°C 1000 hours)
· Rubycon (YK) 200V-100uf 16X25 (400ma Ripple - 85°C 2000 hours)
High Case Profile
24 and 36 mosfet XieChang controllers (Max Vertical Caps 18X40)
· Rubycon (YK) 100V-1000uf 18X40 (1350ma Ripple - 85°C 2000 hours)
· Nichicon (VY) 160V-470uf 18X40 (855ma Ripple - 105°C 1000 hours)
· Rubycon (YK) 160V-330uf 18X35.5 (760ma Ripple - 85°C 2000 hours)
· Rubycon (YK) 200V-330uf 18X35.5 (680ma Ripple - 85°C 2000 hours)
· Nichicon (VY) 200V-330uf 18X40 (675ma Ripple - 105°C 1000 hours)
The reported ripple current relates to the ESR ... discussion welcome.
pdf of Doc's Cap reference here ... http://98.131.176.65/endless-sphere/ABA0000CE33.pdfDoctorbass said:
Unfortunately 22x40 caps really don't fit on the 24-mosfet board ... or maybe they will ... Hmmm.
steveo
100 kW
Hey Knuckles
you wouldn't have the specs for the 200v 470uf would you?
i will try to get them on my end..
-steveo
you wouldn't have the specs for the 200v 470uf would you?
i will try to get them on my end..
-steveo
Knuckles said:GENERAL PURPOSE (cheapo)
ALUMINUM ELECTROLYTIC CAPACITORS
Standard Case Profile
6, 12 and 18 mosfet XieChang controllers (Max Vertical Caps 16X25)
· Nichicon (VK) 100V-470uf 16X25 (1000ma Ripple - 85°C 1000 hours)
· Rubycon (YK) 100V-470uf 16X25 (950ma Ripple - 85°C 2000 hours)
· Nichicon (VZ) 100V-470uf 16X25 (715ma Ripple - 105°C 1000 hours)
· Nichicon (VY) 100V-470uf 16X25 (715ma Ripple - 105°C 1000 hours)
· Nichicon (VK) 200V-100uf 16X25 (450ma Ripple - 85°C 1000 hours)
· Rubycon (YK) 200V-100uf 16X25 (400ma Ripple - 85°C 2000 hours)
High Case Profile
24 and 36 mosfet XieChang controllers (Max Vertical Caps 18X40)
· Rubycon (YK) 100V-1000uf 18X40 (1350ma Ripple - 85°C 2000 hours)
· Nichicon (VY) 160V-470uf 18X40 (855ma Ripple - 105°C 1000 hours)
· Rubycon (YK) 160V-330uf 18X35.5 (760ma Ripple - 85°C 2000 hours)
· Rubycon (YK) 200V-330uf 18X35.5 (680ma Ripple - 85°C 2000 hours)
· Nichicon (VY) 200V-330uf 18X40 (675ma Ripple - 105°C 1000 hours)
The reported ripple current relates to the ESR ... discussion welcome.
pdf of Doc's Cap reference here ... http://98.131.176.65/endless-sphere/ABA0000CE33.pdfDoctorbass said:
Unfortunately 22x40 caps really don't fit on the 24-mosfet board ... or maybe they will ... Hmmm.
Knuckles
10 kW
I didn't come across it. That's why I was surprised with the 200V 470uf value.steveo said:
I'll keep looking. I assume these are Rubycon but the series value makes a difference of course.
If I was to guess I'd say maybe you have 6 - 200V-470uf 18X40 800ma Ripple capacitors.
Is this a good thing?
LIHAFChttp://98.131.176.65/endless-sphere/A0A0A86888CE4C0591627D18FF09455A.pdf
Knuckles
10 kW
New 24-Fet Voltage Regulator Circuit
The new 24-fet controller board has a nicely improved voltage regulator circuit that appears superior to the VR circuit on the older 18-fet XieChang board.
Looking at the new VR circuit it is clear that it is a simple "buck" step down converter tailored to the specific requirements of the controller board 5 volt and 15 volt circuitry. It should be very efficient (generating very little waste heat) at dropping the battery voltage from 100V (max) down to the input of the 15V linear regulator. The voltage “window†for safe controller operation should be very large easily accepting any battery from 24 to 72 volts with no modifications needed.
This is my best guess (as of today) of the functional parts of the new switching regulator circuit.
Higher res pic
How I Think It Works
I don’t physically have this board (yet) so I am guessing how the VR circuit might work (based on the pics from Keywin). Comments and/or corrections are welcome of course.
BTW, I like to call the “thin red wire†that connects the VR circuit to the battery (+) the “ignition wire†as it is meant to be used with a key switch to turn the controller on and off. Also, there is space for a single “power resistor†(R115) that may well allow for voltages in excess of 100V max. For now, let’s assume R115 is zero ohms (a piece of wire) when analyzing the VR circuit. I’ll try guessing at values for R115 later on.
FYI, the “7550-1†is a 5V linear regulator and the “78L15†is a 15V linear regulator with a rated max input voltage of 30V. The “100V diode†is a SS110 (100V) schottky diode that (I believe) conducts current from GND during the “buck†switching cycles. I think that Q13 is the “buck†switch (I’m guessing it’s a mosfet) and that Q17 and Q18 are transistors that create a hysteresis feed back loop that controls Q13 switching frequency and keeps the input to the 15V regulator below the 30V max value. Also, my experience with 12-fet controllers tells me that the minimum current to the 15V regulator is about 30ma (motor, halls and throttle not connected) AND maybe 70ma (everything hooked up with motor cruising under load at 50% wide open throttle). The 15V bus powers the gate drivers (24 mosfets) and the 5V bus powers the “116†chip and the throttle/motor hall sensors.
I am guessing that this “buck†converter is switching at a very high but varying harmonic frequency depending on the loads and voltages applied to the circuit. Again, I am just guessing here. So let’s slow down time into microseconds and try to see what’s going on.
Idea 1: Ignition off. All voltages and currents in the circuit are ZERO. Q13 (switch) is OPEN. Transistors Q17 and Q18 are open (not conducting any current).
Idea 2: Ignition just turned on (say 100V bat). Cap C2 starts to charge up and the voltage increases. R117(15k) and R118(3.1k) start to conduct (this is a voltage divider). At 100V bat this divider (midpoint) cannot exceed 17V. The current thru this divider cannot exceed 5.5ma.
Idea 3: The divider is connected to the base of Q17. Q17 is placed between 2 resistors (R119 and R120, both 2.1k). Before Q17 starts to conduct, the voltage across it is full bat voltage 100V.
Idea 4: As the base voltage of Q17 begins to rise, Q17 starts to conduct. AND this circuit “pulls down†the gate voltage of Q13 (and current starts to conduct thru R116 (1.2k)). Q13 is now CLOSED and current runs to the “buck†inductor. The voltage across the inductor “jumps†from zero to 100V and current starts flowing thru the inductor and the inductor magnetic field grows (storing energy).
Idea 5: The more Q17 conducts, the lower the voltage drop across it. If Q17 was completely closed the voltage across it would be close to zero. Now we would have three resistors in series (R116, R119 and R120). But Q17 emitter voltage would be about 39V way above the base voltage max of 17V (Q17 would stop conducting). A-Ha! But we do have (a very small) cap C129 in series with R121. This provides some hysteresis causing a momentary voltage boost above 17V. This is a funky circuit. :lol:
Idea 6: As the voltage to the 15V regulator climbs to 30V (max allowable), ANOTHER voltage divider (R124-3.3k and R123-2.2k) is conducting. Eventually the voltage (in) to the 15V regulator climbs to maybe 30V (max allowable). When this happens (I guess) the Zener diode (ZD1) starts to conduct. AND this causes the base of Q18 to rise thus causing Q18 to conduct. When Q18 conducts the base of Q17 plummets to zero volts. And this OPENS switch Q13.
Idea 7: Well the inductor is “energized†but Q13 is open so the voltage before the inductor MUST drop BELOW zero volts as it “pumps†current (releases energy) into the 30V cap that feeds the 15V regulator. The schottky diode completes the circuit and conducts current from ground to the inductor coil.
Idea 8: Eventually the voltage to the 15V regulator drops under the 30V (max allowable), and Zener diode (ZD1) stops conducting. The base of Q18 goes to zero and the Q17 - Q19 “dance†starts up again (charging) the inductor and maintaining the 30V “bus†to the 15V regulator.
Cool. 8)
The new 24-fet controller board has a nicely improved voltage regulator circuit that appears superior to the VR circuit on the older 18-fet XieChang board.
Looking at the new VR circuit it is clear that it is a simple "buck" step down converter tailored to the specific requirements of the controller board 5 volt and 15 volt circuitry. It should be very efficient (generating very little waste heat) at dropping the battery voltage from 100V (max) down to the input of the 15V linear regulator. The voltage “window†for safe controller operation should be very large easily accepting any battery from 24 to 72 volts with no modifications needed.
This is my best guess (as of today) of the functional parts of the new switching regulator circuit.
Higher res picHow I Think It Works
I don’t physically have this board (yet) so I am guessing how the VR circuit might work (based on the pics from Keywin). Comments and/or corrections are welcome of course.
BTW, I like to call the “thin red wire†that connects the VR circuit to the battery (+) the “ignition wire†as it is meant to be used with a key switch to turn the controller on and off. Also, there is space for a single “power resistor†(R115) that may well allow for voltages in excess of 100V max. For now, let’s assume R115 is zero ohms (a piece of wire) when analyzing the VR circuit. I’ll try guessing at values for R115 later on.
FYI, the “7550-1†is a 5V linear regulator and the “78L15†is a 15V linear regulator with a rated max input voltage of 30V. The “100V diode†is a SS110 (100V) schottky diode that (I believe) conducts current from GND during the “buck†switching cycles. I think that Q13 is the “buck†switch (I’m guessing it’s a mosfet) and that Q17 and Q18 are transistors that create a hysteresis feed back loop that controls Q13 switching frequency and keeps the input to the 15V regulator below the 30V max value. Also, my experience with 12-fet controllers tells me that the minimum current to the 15V regulator is about 30ma (motor, halls and throttle not connected) AND maybe 70ma (everything hooked up with motor cruising under load at 50% wide open throttle). The 15V bus powers the gate drivers (24 mosfets) and the 5V bus powers the “116†chip and the throttle/motor hall sensors.
I am guessing that this “buck†converter is switching at a very high but varying harmonic frequency depending on the loads and voltages applied to the circuit. Again, I am just guessing here. So let’s slow down time into microseconds and try to see what’s going on.
Idea 1: Ignition off. All voltages and currents in the circuit are ZERO. Q13 (switch) is OPEN. Transistors Q17 and Q18 are open (not conducting any current).
Idea 2: Ignition just turned on (say 100V bat). Cap C2 starts to charge up and the voltage increases. R117(15k) and R118(3.1k) start to conduct (this is a voltage divider). At 100V bat this divider (midpoint) cannot exceed 17V. The current thru this divider cannot exceed 5.5ma.
Idea 3: The divider is connected to the base of Q17. Q17 is placed between 2 resistors (R119 and R120, both 2.1k). Before Q17 starts to conduct, the voltage across it is full bat voltage 100V.
Idea 4: As the base voltage of Q17 begins to rise, Q17 starts to conduct. AND this circuit “pulls down†the gate voltage of Q13 (and current starts to conduct thru R116 (1.2k)). Q13 is now CLOSED and current runs to the “buck†inductor. The voltage across the inductor “jumps†from zero to 100V and current starts flowing thru the inductor and the inductor magnetic field grows (storing energy).
Idea 5: The more Q17 conducts, the lower the voltage drop across it. If Q17 was completely closed the voltage across it would be close to zero. Now we would have three resistors in series (R116, R119 and R120). But Q17 emitter voltage would be about 39V way above the base voltage max of 17V (Q17 would stop conducting). A-Ha! But we do have (a very small) cap C129 in series with R121. This provides some hysteresis causing a momentary voltage boost above 17V. This is a funky circuit. :lol:
Idea 6: As the voltage to the 15V regulator climbs to 30V (max allowable), ANOTHER voltage divider (R124-3.3k and R123-2.2k) is conducting. Eventually the voltage (in) to the 15V regulator climbs to maybe 30V (max allowable). When this happens (I guess) the Zener diode (ZD1) starts to conduct. AND this causes the base of Q18 to rise thus causing Q18 to conduct. When Q18 conducts the base of Q17 plummets to zero volts. And this OPENS switch Q13.
Idea 7: Well the inductor is “energized†but Q13 is open so the voltage before the inductor MUST drop BELOW zero volts as it “pumps†current (releases energy) into the 30V cap that feeds the 15V regulator. The schottky diode completes the circuit and conducts current from ground to the inductor coil.
Idea 8: Eventually the voltage to the 15V regulator drops under the 30V (max allowable), and Zener diode (ZD1) stops conducting. The base of Q18 goes to zero and the Q17 - Q19 “dance†starts up again (charging) the inductor and maintaining the 30V “bus†to the 15V regulator.
Cool. 8)
