PWM Regulator-Mag D, JM-PhD-D1, UPDATE

Candle Power Flashlight Forum

Help Support CPF:

petrev said:
Also . . . is it OK to cut a big hole in the KIU to allow full in-situ reprogramming ? ie. Access to the port from above while in the torch ? Too much heat from above ? ? ? Or just a stupid idea ?

Cheers
Pete
Click to expand...
Sure, you can do this. But I'd suggest using some insulation over an aluminum sheetmetal donut style cover. While in use, the insulation will hold the donut in place and isolate the "gut". The programming plug is black, so if it ever had direct exposure to the bulb while on, it would melt after some time.
 
JimmyM said:
Sure, you can do this. But I'd suggest using some insulation over an aluminum sheetmetal donut style cover. While in use, the insulation will hold the donut in place and isolate the "gut". The programming plug is black, so if it ever had direct exposure to the bulb while on, it would melt after some time.
Click to expand...

Thanks Jimmy

Nice idea

:thumbsup:
 
Jimmy, if I need an RMS meter, how about the fluke 87 series III. They seem to be cheaper than the newer models.
 
Starlight said:
Jimmy, if I need an RMS meter, how about the fluke 87 series III. They seem to be cheaper than the newer models.
Click to expand...
It has to say True RMS. I'm not sure about the 87III. You might want to email Fluke and ask them about square wave and PWM signals.
If you can find one, get the Fluke 189. It's more expensive, but worth EVERY PENNY if you do anything with non-DC voltages. I was thinking about picking up another. But haven't had the extra $$$ lately.
 
I checked some more on the meters. Most of Fluke's meters say they measure true rms, but only the 187, 189, 287, and 289 seem to have the AC + DC function, when measuring voltage or current. It seems that several people have said that this summing function is what we need to be measuring to set the regulators.
 
It is not enough to be True RMS. The meeter has to have the function/capability of measuring a signal with a duty cycle away from the normal/typical 50%. There might be other meters/brands that have this capability, but we know hands-on that the 4 Fluke models listed above have the "AC+DC" function that correctly measures the PWM signal generated by the PhD drivers.
 
The 187 is the cheapest of those listed above.
Will, is right. True RMS is not enough. The AC+DC mode is able to measure the equivalent of an AC wave form with a DC offset. Which is pretty much what our PWM is.
Search eBay for "Fluke 187"
 
JimmyM said:
The 187 is the cheapest of those listed above.
Will, is right. True RMS is not enough. The AC+DC mode is able to measure the equivalent of an AC wave form with a DC offset. Which is pretty much what our PWM is.
Search eBay for "Fluke 187"
Click to expand...

Just to confirm again, my 179 says "True RMS" but no AC + DC capability. I ended up getting a 189 on EBay for a great price (under $200). Pristine condition. Look at the photos, and details...for example this is a good one if price stays low.
 
JimmyM said:
Excellent, Pete. Let us know how it does.
You're right. the 189 rules. However, I still want the 287/289
Click to expand...

287/289 = Mucho Dinero

but NICE

Fitted a little 5V switch into my AVRISP programmer

Drilled my KIU adapter but the provided template didn't quite match my JM-PHD-D1 so I have made a new one using photos and adjusted to fit my 2 samples - just the tiniest variation in the assembly process.

PDF here .

Feel free to copy it to your host if you want to let people use it as an alternative template.

Re-Drilling my Mag Switch body to fit JM-PhD (slightly different to my FET switch of course)

. . .

Cheers
Pete
 
petrev said:
287/289 = Mucho Dinero

but NICE

Fitted a little 5V switch into my AVRISP programmer

Drilled my KIU adapter but the provided template didn't quite match my JM-PHD-D1 so I have made a new one using photos and adjusted to fit my 2 samples - just the tiniest variation in the assembly process.

PDF here .

Feel free to copy it to your host if you want to let people use it as an alternative template.

Re-Drilling my Mag Switch body to fit JM-PhD (slightly different to my FET switch of course)

. . .

Cheers
Pete
Click to expand...
Thanks, Pete. I'll use yours. That's wierd though. I made that drill template from the PCB design.
Oh well.
Yeah, mucho dinero is right. But sooooo pretty.
 
Hi Jimmy

Fitted everything

Using the 64333 set
1. 64633(6xA123) - Set Vbulb 16.8 Vlow 17.4

Checked with a 64458 and the 6xA123

Steady 16.68V AC+DC

but noticed occasional overshoot so switched to MAX and found

17.48V MAX 16.68V Steady Regulation
(upto 17.63 max variable not every time.)

and occasional
variable up to 17.815 MAX spike during switch off

Went ahead and fitted a 64333 in the hope that this was just a very brief transient that would not cause problems
but 64333 intstaflash.😳

Let me know what you think . . .

Cheers
Pete
 
Last edited:
JimmyM said:
Thanks, Pete. I'll use yours. That's wierd though. I made that drill template from the PCB design.
Oh well.
Yeah, mucho dinero is right. But sooooo pretty.
Click to expand...

Mmmmmm So Pretty - and useful too.

Did you use the middle of the component position on the PCB layout ? Depends what size drill people are using too I guess and the little screws are right in the corner.

I am using 3mm drill holes and I managed to get a bit of drift with one of them so the little screw is right at the edge again ! Ah well.

Better check mine is right by your eye too.

Cheers
Pete
 
petrev said:
Hi Jimmy

Fitted everything

Using the 64333 set
1. 64633(6xA123) - Set Vbulb 16.8 Vlow 17.4

Checked with a 64458 and the 6xA123

Steady 16.68V AC+DC

but noticed occasional overshoot so switched to MAX and found

17.48V MAX 16.68V Steady Regulation
(upto 17.63 max variable not every time.)

and occasional
variable up to 17.815 MAX spike during switch off

Went ahead and fitted a 64333 in the hope that this was just a very brief transient that would not cause problems
but 64333 intstaflash.😳

Let me know what you think . . .

Cheers
Pete
Click to expand...
During softstart it ramps in 5 PWM step (each step is 0.09775%) increments until it exceeds computed output, then ramps back to steady state in 1 PWM step increments. However, the steps happen 244 times per second. So it's only 5 steps above (Maximum. It could between 1 thru 4) the set point for 1/244th of a second (One pulse). 5 steps is 0.2V. Then it can take as long as 5/244th of a second to reach equilibrium. That's only 20 mS.
Keep in mind that the Fluke does take time to settle in at it's RMS voltage. On a scope, you can't see the the overshoot because the overshoot is only a few pulses. If your steady state loaded RMS voltage is 16.68V, then the most it could overshoot by is 0.20V and that's too quick for the Fluke to capture in an accurate manner.
Do you have your AVR programmer set up yet?
 
Hi Jimmy

I have downloaded and installed Studio and WinAVR and installed the USB driver in the process . . .

I haven't tried any actual programming yet ! ! ! but I have your step by step guide.

On the measurements front - it is not consistent - the overshoot on power up happens maybe 1 in 3 and the spike before switch off similarly but those are the readings from the old Fluke with the peak hold(MAX) function to capture what I noticed first by eye ! and the bulb did flash.

Maybe a faulty bulb but I don't want to risk another without checking it all with you.

Cheers
Pete

ps. Can't find your bannerad - but your sig line looks good.
 
petrev said:
Hi Jimmy

I have downloaded and installed Studio and WinAVR and installed the USB driver in the process . . .

I haven't tried any actual programming yet ! ! ! but I have your step by step guide.

On the measurements front - it is not consistent - the overshoot on power up happens maybe 1 in 3 and the spike before switch off similarly but those are the readings from the old Fluke with the peak hold(MAX) function to capture what I noticed first by eye ! and the bulb did flash.

Maybe a faulty bulb but I don't want to risk another without checking it all with you.

Cheers
Pete

ps. Can't find your bannerad - but your sig line looks good.
Click to expand...

I just looked at the code. It is as I described above. Let me work on a couple of things. I may send you a piece of experimental software to try out. I still don't think that it's the software. I've tested 64275s at 7.15V with a 40V source. If that overshoot existed as an overall design flaw it would blown that bulb sure enough.
Let me work for a bit.

The banner just isn't in rotation yet.
 
Jimmy, I had some voltage setting problems too. You sent me one regulator set up for 18v output and 26v low battery. I put it in a light with 7 emoli cells and a 64458 bulb, but the low voltage pulsing occurred. I then put in an 8th cell. Pack voltage was 32.4v but it still pulsed. I had the turn the pot more than one turn to get the pulsing to stop. There is no way that 10 amps should take a 32v pack below 26v. I intend to get a meter soon to check voltages myself.
 
Starlight said:
Jimmy, I had some voltage setting problems too. You sent me one regulator set up for 18v output and 26v low battery. I put it in a light with 7 emoli cells and a 64458 bulb, but the low voltage pulsing occurred. I then put in an 8th cell. Pack voltage was 32.4v but it still pulsed. I had the turn the pot more than one turn to get the pulsing to stop. There is no way that 10 amps should take a 32v pack below 26v. I intend to get a meter soon to check voltages myself.
Click to expand...
Yes, please do. If the setting is off and you can't seem to correct it to your satisfaction, I'll cross-ship you another and refund you the postage.
I'll cross ship you another, in any case, if you want.
 
I'll be shipping all new regulator purchases with updated software.
Upgrades include:
- Faster response to low voltage and over temp conditions. 0.25 seconds versus 0.5 seconds.
- After the low voltage mode (pulsing) has been in place for 30 seconds, it will shutdown and go to sleep.
- Over temp mode will reduce bulb voltage to 35% of the set voltage instead of 25%
- The higher softstart ramp will disengate before reaching regulation point to eliminate any overshoot of more than 0.04V.
- More adaptive battery voltage sampling. It will average 4 or 8 readings per cycle depending on duty cycle.
- Multiple averaged readings of Vbulb and Vlow pots.

I'll be posting the updated software soon for those who want to re-program theirs.
 
You must log in or register to reply here.

Latest posts

Back
Top