SALE: JM-PhD-D1 PWM Hotwire Regulator

[JimmyM]

7.2V will be hot. Lux's charts show an estimated life of 11.2 hours at 7.2V. A medium soft start would probably be better. With variations in bulb manufacture, a fast start might pop the statistically weaker bulbs.

JimmyM,

What do you think of this configuration for one of the five boards:

2D Mag
3xIMR26500 (using a Mag tail spring mod)
Osram 64250

Operation Mode: Standard
Vbulb: 7.2V
Vlow: 9.0V
Softstart: Fast
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 70C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

[Justin Case]

Rats. I had hoped I could use basically the same settings that seemed to be ok for a WA1111 in Post #373.

If I want to keep the fast start for the 64250, what is the max Vbulb that you'd recommend? And if Fast soft start isn't a good idea for the WA1111 either, what Vbulb would you suggest?

Thanks.

[Linger]

What does this option mean?
"Response: 1/8, 1/4, 1/2 second"
I've read through intro post several times and either its not specified or I keep missing it.

[Justin Case]

In the ATMEL source code, it looks like response is the duration a voltage event must happen to trigger an action such as enable the low voltage mode or over temp mode.

In Post #1, there is also this:

"Pot 2 (Vlow): Adjusts low voltage set point. (Clockwise increases) When the low voltage set point is reached (by falling pack voltage) and maintained for 0.125 to 0.5 seconds (configurable by the user), the output will be reduced and ramped up and down between 2 lower percentages of the set output voltage. It will pulse. If the batteries recover, the regulator will ramp voltage back up to the set point."

That seems consistent with what is commented in the source code.

[JimmyM]

In the ATMEL source code, it looks like response is the duration a voltage event must happen to trigger an action such as enable the low voltage mode or over temp mode.

Right on the money, Justin. It just makes sure 1 or 2 high/low readings don't prematurely trigger an event. There's a lot of switching noise due to the large current transients.

Re: WA1164. You could try the fast start on a slightly lower voltage and move up to where you want to be. But with bulb life in the 10 hour area, I can't make any promises.

[Justin Case]

JimmyM, can you help me with differentiating between the various soft start speeds.

I see in the code that slow is about 1 sec, med is about 1/2 sec, and fast is about 1/4 sec. Your Post #1 also says: "S-Curve softstart ramp: This function is not an option, it's more of a feature. It starts the bulb slowly at first, then ramps quickly to a point less than the regulated voltage then ramps slowly to the target voltage. It's designed to improve bulb care and hopefully blow up fewer bulbs."

I assume by s-curve, the soft start looks like this:
-
-
-
-
-
-
-
-
-
-

I guess what I'm unsure of is if the time to reach say the knee of the s-curve is the same for all 3 speeds (call it Tknee), with the remaining time spent in the slow ramp to the target voltage (1sec - Tknee, 1/2sec - Tknee, or 1/4sec - Tknee).

Or does the quick ramp up also have different slopes for the 3 soft start speeds?

Visually, it's hard for me to really see much, if any difference in Fast vs Medium soft start. They both seem to get the beam up to speed equally as fast. Granted, my eyeball is probably not the greatest measurement device in the world. But that's why I was wondering if the time I called Tknee above might be the same for all 3 soft start speeds.

[JimmyM]

There are 3 parts to the "S". It's not really a true S curve. It's more like 3 separate slopes.
(bear with me, this is from memory)
For SLOW. ~1 second. The first part of the slope is shallow and ends at ~ 1/4 Vbulb, then it steepens and continues until 7/8th of Vbulb. Then shallows again until Vbulb is reached.

For MEDIUM. ~1/2 second. The first part of the slope is shallow and ends at ~ 1/8 Vbulb, then it steepens (steeper than SLOW) and continues until 15/16th of Vbulb. Then shallows again until Vbulb is reached.

for FAST. ~100-200ms. There is no "first part of the slope". It's very steep, and continues until 15/16th of Vbulb. Then shallows again until Vbulb is reached.

Not much difference to the eye between FAST and MEDIUM, but the Bulb will notice. SLOW is really only for pushing the big bulbs VERY hard since it creeps up on Vbat more slowly for longer in the 3rd slope where the real overdrive voltage is put in. 7/8th of Vbulb. For a 64458, that means the 3rd slope slowly ramps from ~17.5V to 20V. You can direct drive one on 17.5V so there's no danger of the filament blowing, but easing up to 20 or even 21V is critical if you want the bulb to last more than 1 or 2 starts.

JimmyM, can you help me with differentiating between the various soft start speeds.

I see in the code that slow is about 1 sec, med is about 1/2 sec, and fast is about 1/4 sec. Your Post #1 also says: "S-Curve softstart ramp: This function is not an option, it's more of a feature. It starts the bulb slowly at first, then ramps quickly to a point less than the regulated voltage then ramps slowly to the target voltage. It's designed to improve bulb care and hopefully blow up fewer bulbs."

I assume by s-curve, the soft start looks like this:


I guess what I'm unsure of is if the time to reach say the knee of the s-curve is the same for all 3 speeds (call it Tknee), with the remaining time spent in the slow ramp to the target voltage (1sec - Tknee, 1/2sec - Tknee, or 1/4sec - Tknee).

Or does the quick ramp up also have different slopes for the 3 soft start speeds?

Visually, it's hard for me to really see much, if any difference in Fast vs Medium soft start. They both seem to get the beam up to speed equally as fast. Granted, my eyeball is probably not the greatest measurement device in the world. But that's why I was wondering if the time I called Tknee above might be the same for all 3 soft start speeds.

[Justin Case]

Thanks. It seems that in terms of perceived brightness ramp-up speed perhaps there is little need for selecting Fast soft start over Medium, while Medium offers advantages over Fast in terms of driving a filament closer to its "flash point".

[JimmyM]

Thanks. It seems that in terms of perceived brightness ramp-up speed perhaps there is little need for selecting Fast soft start over Medium, while Medium offers advantages over Fast in terms of driving a filament closer to its "flash point".

Indeed. I suggest Medium for all lights except in situations I described above. However, since it's possible and configurable, I made a Fast option for those who feel they need as fast a start as possible. Alan's start is in the 100-200ms range if memory serves.

[Justin Case]

A followup on the soft start ramp-up. How do the 3 speeds compare to the "natural" voltage ramp-up one would get from direct drive from say 3xIMR26500 cells that were sufficiently drained so that their resting voltage was 11.8V, 12.0V, or whatever might be needed to avoid flashing a Hikari JC-5607?

[Justin Case]

JimmyM,

Here are my selections for the five regulators. #1 to #3 are identical, and duplicate my previous order for a 15.5V Mag623 regulator. #4 is meant to drive a compact Maglite, while balancing output, run time, and bulb life. #5 duplicates Linger's setup. Please let me know if you spot any problem choices.

Thanks.

#1
4D Mag623
5xAW IMR26500 cells

Operation Mode: Standard
Vbulb: 15.5
Vlow: 15.0
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 80C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#2
4D Mag623
5xAW IMR26500 cells

Operation Mode: Standard
Vbulb: 15.5
Vlow: 15.0
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 80C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#3
4D Mag623
5xAW IMR26500 cells

Operation Mode: Standard
Vbulb: 15.5
Vlow: 15.0
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 80C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#4
2D Mag
3xAW IMR26500 (using a Mag tail spring mod)
Osram 64250

Operation Mode: Standard
Vbulb: 7.2V
Vlow: 9.0V
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 70C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#5
Multi-Level WA 1164/JC5607 x 3s IMR 26500/26650
Vbulb: 9.5v
Vlow: 9.6v

Level-1: 18-100% (70%)-6volts
Level-2: 18-100% (85%)
Level-3: 18-100% (100%)

SoftRamp: Yes
Softstart: Slow
Low Volt Warning: Pulsing
Low volt shutdown: Both (30s/94%) just going with default suggestion here
Over Temp trigger: 50C, 60C, 70C, or 80C
Over Temp Shutdown: Yes/No, If, Yes, Y seconds: (15 Seconds)
Response: 1/8, 1/4, 1/2 second

[JimmyM]

Everything looks pretty straight forward below. I'll see if I can get them programmed tonight.

JimmyM,

Here are my selections for the five regulators. #1 to #3 are identical, and duplicate my previous order for a 15.5V Mag623 regulator. #4 is meant to drive a compact Maglite, while balancing output, run time, and bulb life. #5 duplicates Linger's setup. Please let me know if you spot any problem choices.

Thanks.

#1
4D Mag623
5xAW IMR26500 cells

Operation Mode: Standard
Vbulb: 15.5
Vlow: 15.0
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 80C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#2
4D Mag623
5xAW IMR26500 cells

Operation Mode: Standard
Vbulb: 15.5
Vlow: 15.0
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 80C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#3
4D Mag623
5xAW IMR26500 cells

Operation Mode: Standard
Vbulb: 15.5
Vlow: 15.0
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 80C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#4
2D Mag
3xAW IMR26500 (using a Mag tail spring mod)
Osram 64250

Operation Mode: Standard
Vbulb: 7.2V
Vlow: 9.0V
Softstart: Medium
Low Volt Warning: Pulse
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 70C
Over Temp Shutdown: Yes, 30 sec
Response: 1/4 sec

#5
Multi-Level WA 1164/JC5607 x 3s IMR 26500/26650
Vbulb: 9.5v
Vlow: 9.6v

Level-1: 18-100% (70%)-6volts
Level-2: 18-100% (85%)
Level-3: 18-100% (100%)

SoftRamp: Yes
Softstart: Slow
Low Volt Warning: Pulsing
Low volt shutdown: Both (30s/94%) just going with default suggestion here
Over Temp trigger: 50C, 60C, 70C, or 80C
Over Temp Shutdown: Yes/No, If, Yes, Y seconds: (15 Seconds)
Response: 1/8, 1/4, 1/2 second

[JimmyM]

Programmed, packed, labeled and should ship tomorrow AM.

[JimmyM]

I was just looking again at this great product.
I can only tell that these regulators are the best thing one can have with incans. There simply must be a next run because I did not buy enough of them for future needs.

On a next run could there be a optional extension cable to reprogram the light while still under the kiu socket?
If the pulsing/ pwm frequency would be higher and the minimum required battery voltage lower (3,0 volt) would it also be usable for LED use?
That would rock especially with the new 3,6 volt 9 amp 30 watt leds.

I think I've found an FET that will do the job for LED use.
IRFR3706 by International Rectifier. It is good up to 30V/65Amax. But the inportant thing is the low voltage performance. It should carry 9A with no problems with a 3V gate. Of course testing would be required.

[^^Nova^^]

Jimmy,

How would a single mode one of these cope if it was wired permanently on and used with a tail switch?

Cheers,
Nova

[JimmyM]

Jimmy,

How would a single mode one of these cope if it was wired permanently on and used with a tail switch?

Cheers,
Nova

It should be fine. Of course the tailswitch will have to carry the full current of the bulb, but if that's OK, then the regulator will work as expected.
In the "normal" wiring method, you will still have the FET leakange current of a few nano-amps. With a tailswitch, there will be zero leakage.

[JimmyM]

Another furum member was having repeated problems with his JM-SST softstarter (51V, 400W bulb). So after having it shipped back to me for the third time +/-, rework had totally worn out the board. So I took it as an opportunity for experimentation and redesign. I based the new softstarter on a microcontroller (Tiny45 8TSSOP package). I know how to use them now, why not. But wait, if I added 2 resistors, it could become a regulator instead of just a softstarter. So the resulting design I have has a primary regulator that steps down as much as 80V to 11.9V. That feeds a 5V regulator to power the microcontroller. The microcontroller drives a MOSFET driver that switches the 75V D2PAK FET gate with 11.9V@>2A. So it switches really fast. There are no adjustment pots. No room.
I tested it tonight. A set voltage (in software) of 20.0V equated to 20.06V measured. I haven't done high power testing yet. I have another 75V FET to try as well.
After some more testing I'll post results and pictures in a new thread. This may lead to a new design. The PhD-JM-D2. But we'll see.

[petrev]

One of my SSTs is still doing sterling service at 33V and 700-1100W (Twin outboard FETs)

As always looking forward to your experiments . . .

Cheers
Pete

[JimmyM]

One of my SSTs is still doing sterling service at 33V and 700-1100W (Twin outboard FETs)

As always looking forward to your experiments . . .

Cheers
Pete

I'll probably send you the prototype X1 for full power testing since you already have a big light and have re-programmed my regulators before.

[petrev]

[JimmyM]

Here's a pic of that custom regulator I built.

[LumenHound]

Great Ceasar's Ghost!

Jimmy that looks outstanding.

[hammockjockey]

1 driver please

Operation Mode: Standard
Vbulb: 12 volts for 1185 bulb
Vlow: 10 volts
Softstart: Slow
Low Volt Warning: Pulsing
Low volt shutdown: Both. Timed:30 seconds, Undervolt: 94%
Over Temp trigger: 70C
Over Temp Shutdown: Yes, 30 Seconds
Response: 1/4 sec

What would you suggest for my application for the vlow and undervolt values? I'm using 3 26650 IMR 4000mah 3.7 volt cells and want to make sure I don't overdischarge them. Thanks for your advice!

paypal of $47 sent!

[JimmyM]

1 driver please
paypal of $47 sent!

I'm sorry. There are sold out. I forgot to change the original post to indicate that they are sold out. I've done that just now.
I'll bounce your PayPal back.

[donn_]

Jimmy, what's the ETA on more stock?

[JimmyM]

Jimmy, what's the ETA on more stock?

I don't know just yet. I have a few other projects in the works. I'm moving forward with the X1. It's a low volume design just for the Big lantern Larry14K/PK1000 guys. 40+ amps. There may be a "D2" regulator that is programmable only. No pots. But it has greater power handling. >20A possibly.
So it may be several months. When it's coming. I'll post here and start a new sales thread.

[donn_]

Curses! I didn't act early enough. I have a project going which involves using a Navy Battle Lantern with 8S2P 26650 A123 cells driving a big incan through a flood lens.

If you dig anything up which will work, please let me know.

[Roland]

This might be a nice LED for a flashlight which needs regulation:

Luminus CSM-360-W
http://www.luminus.com/stuff/content...lumination.pdf
Standard specifications:
over 6,000 lumens
1A to 6.3A at 12V to 14V
Maximum operating temperature 150 degrees celcius
integrated temperature sensor in the led that can be red out electronically.

It would be about the light output of a hard driven 623 in a long-life drop proof led with about half the power consumption which means longer runtimes. (when the right regulator exists)

[JimmyM]

Curses! I didn't act early enough. I have a project going which involves using a Navy Battle Lantern with 8S2P 26650 A123 cells driving a big incan through a flood lens.

If you dig anything up which will work, please let me know.

I love projects like that.
I'm currently working on the X1 design and am in the final stages. I need to clean up the board design and build a proper bill of materials to estimate costs more accurately. I can't build them until I have finances worked out on my end.

[^^Nova^^]

I'll be up for a couple when you do another run.

These are simply the best.

Cheers,
Nova