The old 37V to 15V conversion question again

[Reid Welch]

double post in error

 

[Reid Welch]

Matt: the controller as converter idea proved topheavy here;
I didn't find it stable-it's unregulatable as the battery supply voltage shifts during discharge.

eP, thanks! Complex stuff. Good to know geekdom thrives also in Poland
like a tree grows in Brooklyn (obtuse old American humor). Thanks friend.

Bob! Happy for Rick. Knowing he has a new calling calling, may speed his recovery. Put me down for a converter, Rick.
And if Rick isn't up to it (boy, I need one in two weeks or so when my bike should be back together), then a parts kit would be super convenient, Bob.

I have tons of small resistors, all values, and can find or make a small box, etc.


Thanks Bob, give Rick our best wishes.
Tell him one wisher has ulterior motives (that'd be me).

grins,
R.

 

[fechter]

Matt: the controller as converter idea proved topheavy here;
I didn't find it stable-it's unregulatable as the battery supply voltage shifts during discharge.

It wouldn't be that hard to add a regulator of some sort. For the power, it might be the least expensive option. Nicely packaged to start with also.
I'll have to think about an easy way to regulate one of those.

 

[bobmcree]

using a brushed speed control which has as its heart a pwm chip seems a sort of rube goldberg way to achieve the goal of pwm switching current to a light bulb, and in any case it would be an open loop design and just a dumb way to do it IMO. if you really want to reinvent the mousetrap you can use a 555 as a pwm controller and switch a fet and coil cap diode loop setup, then use voltage mode feedback to achieve a stable output. or for $4 you can use the LM2576 which was designed by National engineers to do exactly this job, has all the necessary hardware to do the job more efficiently, more reliably, smaller, cheaper, and need i say, better?

-bob

 

[fechter]

Of course you're right Bob.
But if you have to order all the parts one off and build a case, solder everything together, it could wind up being nearly as expensive as a cheap chinese controller.

What I was suggesting is that you could make a cheap motor controller closed loop with voltage feedback (regulated). If you needed 10-15 amps output, it might be the cheapest way to go.

For 4 amps, the simple switcher is very nice. Very compact also.

 

[bobmcree]

Of course you're right Bob.

i wonder if you could teach my wife these words? i agree that if you need to control a few hundred watts a brushed speed control is a very cost effective way to do it, and if you wanted to control 4 auto headlights from a 48v ebike battery without connecting them in series this would be a cost effictive solution.

to control a single load of under 50w the simple switcher series by national are a quantum step forward in simplicity of design for an efficient low volume power controller that can easily achieve an output line/load voltage regulation of much better than a percent and requires only a couple of low esr caps and a high q inductor that will not saturate at the required current. Typical caps and inductors are $1-2 in small volume for a 50w implementation. 100 pc prices for components to produce a system that would drop 48v-15v@50W should be well under $10 including the LM2576

just a few years ago you needed a specialized ee to design a switching power supply, but now these chips have made it a job for a bright 6th grader. (too bad there do not seem to be very many of these running american companies these days). I have put together simple switcher supplies with inductors and caps out of my junk box that demonstrate that these ICs can handle a huge variety of components.

Switching power supply design was voodoo until just a few years ago for most of us. We can now take advantage of quantum performance level improvements at costs competitive with old school linear regulators, and the National "simple switcher" series is IMO the best place to get started for anyone who wants to use this technology but might have been intimidated in the past by the design calculations. These chips run at several hundred kHz and it is quite surprising how small the inductors and caps required to control a few hundred watts can be.

-bob


-bob

 

[Reid Welch]

Hi Bob,

I will go that route soon. To just -get something for the first runnings of the revamped bike-

I'll make do for a while with the 12V output, $25 unit from Thunderstruck.

If you or Rick K ever get a kit of parts together, ready for sale,
I will buy that.

For now I'll go the ez way, ready made, not good for the light, but it will run the blower fine:

I asked them if the unit will even operate at as low as 33V (where the EV Tech lipo pack cuts off).
The output at 33V will fall to 11.88V, says Brian at ThunderStruck.
Brian says there is "no way" to adjust the output voltage.
Probably it's potted.

Even so, the thing is neat and it's packaged and it's ten amp capable,
not that I need ten amps

Thanks Richard,

This is the best one I've seen around lately:
http://www.thunderstruck-ev.com/dcdc.htm

There's some smaller ones here:
http://www.salvagetech.com/products.asp?cid=42
but most of these don't span the proper voltage range and the biggest ones are 20w.

 

[bobmcree]

i have not been able to find out if rick will be in any shape soon to get his converters back on sale, and i am in the middle of moving, so it is a bad time for me to do it.

most of these converters can be fooled into producing a higher output by connecting a diode in series with the ground lead. then the output will be a volt higher, 2 diodes could give you a couple volts higher output

-bob

 

[Reid Welch]

Bob, I have figured you out:

You are the 8) Porn Star of the DIY Electronics World!

and that's a compliment because

your brain's elect-trons are hot!

thanks!

 

[fechter]

most of these converters can be fooled into producing a higher output by connecting a diode in series with the ground lead. then the output will be a volt higher, 2 diodes could give you a couple volts higher output

-bob

The units are potted, so digging out the epoxy isn't much fun.

I'd recommend the diode trick just to see if it works. The diodes might need to take a fair percentage of the output current, so they should be rated accordingly. If you put 3 or 4 diodes in series, you should be able to boost the voltage to over 14v.

 

[rkosiorek]

most of these converters can be fooled into producing a higher output by connecting a diode in series with the ground lead. then the output will be a volt higher, 2 diodes could give you a couple volts higher output

-bob

The units are potted, so digging out the epoxy isn't much fun.

I'd recommend the diode trick just to see if it works. The diodes might need to take a fair percentage of the output current, so they should be rated accordingly. If you put 3 or 4 diodes in series, you should be able to boost the voltage to over 14v.

most switchers use the ground or V- as a reference only. usually only need to pass 100 mA or less. 1A diodes should be plenty.

rick

 

[Reid Welch]

The illustrated converter is on its way.

Wow, I'm full of hope now that you guys have given me a simple, cheap way to obtain exactly what is wanted.

My hat's off to you all, even if it does not go as predicted.
I'm sure enough now that it will work as you say.

Will report the results here later.

Thank you.

 

[fechter]

most switchers use the ground or V- as a reference only. usually only need to pass 100 mA or less. 1A diodes should be plenty.

rick

That would be true for a linear regulator, but the current from the inductor has to pass through the diode with a switcher. The average current through the diode will be less than the output current, but could be a sizeable percentage of the output under some conditions.

Diodes are cheap. For a 20w load, I'd go with 3A or higher rating.

 

[Doctorbass]

After finfing alot of documents about the famous Dewalt 36V Li-ion from A123, i found somethiong that could interest you guys searching for 37 to 15V.

That 36V battery include a speed controller (PWM type) INSIDE the battery pack. My friend bought one with a flashlight in the kit. that flashlight use a 20V lamp. The way to decrease the 36V (33 real volts) to 20V is by reducing the pulse width (same kind of circuit on page 1 of this topic).

The flashlight connector have a couple of wire: 2 for the input voltage for the lamp, and 3 for a VOLTAGE DIVIDER included inside the flashlight. the resistor value of it are calculated to adjust the PWM inside the pack to 20V.

That battery have a controller that is adjusted by exterior voltage divider depending on the tool that is used with!... that controller can be used with load up to 750W

I know that these controller can be found seperatly on ebay and many RC sites.

I wonder if it cold be used on DC E-bike motor... hummm...

Hope this could help...

Doc

 

[Reid Welch]

Thanks Doc. That'll want some sorting out I'll let others do.

____________________________________


For now, the ThunderStruck DC/DC converter has just arrived in the mail from California. Postal mail is quick.
Thank you Brian Hall of Thunderstruck.

Cost delivered: $29
It's in a good-looking diecast box.

250 grams.


______________________


There is no documentation with the unit, no directions.

The box has three wires, black yellow, red.

Q: What wire goes where,
and does it matter if the unit is powered up without a load?






Off to Radio Shack to get some diodes (see fechter and Bob's posts above).
Also want some Euro Style (grub screw) connectors.

We'll do a lash up to the new EV Tech lipo pack.
Do it atop the washing machine where it's easy to take good pictures.

I hope Maytag doesn't mind

:wink:

 

[Reid Welch]

11.88 V.

The Drain Brain is ahead of the converter.
The lamp is rated 20W at 12V. I have not measured the actual draw.
The losses would seem to be five or six watts within the converter.

A diode in series with the black wire does not boost the voltage--only drops it to 11V and the diode heats.

What do these clues indicate?
I guess it is not a switching type converter--Is it a PWM converter;
in effect, a speed controller with feedback loop as fechter spoke about doing earlier?

Or is the diode to go in series with the yellow lead?
No, that can't work at all.


_____________

Even if this is all I may get from this unit, it works to get me going.
The most important accessory to power is the motor cooling blower.
This will do that job fine and will still provide a usable headlight.
It's just so much better to overvolt the lamp. Incredibly better light.

Thanks for helps again,
Reid

 

[fechter]

If the diode heats, it means there's lots of current going through it.

Too bad it did not increase the output voltage.

Maybe if you send it over to Tyler he can dig all the epoxy out so you can get to the adjustment.

Let me think about it..... Perhaps there's another way.

too bad they don't make 9v lamps.