Sold/Expired Pre-Order JM-PhD-X1, PWM Lantern Regulator

[JimmyM]

I've just received most of the parts (including the ones scheduled to ship in January). I'm waiting on the FETs, Digikey was low on stock and they'll arrive in November. I ordered the boards last night, they'll take about a week.
Once everything arrives, I'll give another update.
Christmas may come early.

 

[Starlight]

Jimmy, great news. A few of those may be mine.

 

[Starlight]

Jimmy,
Since I hope to be first in line for these, I will take a shot at specs for mine. If you need other info, let me know.

NEW X1 REGULATOR #1 for 64663 bulb
Vbulb(max): 39v
Level1 Percentage: 50% - 20v
Level2 Percentage: 69% - 27v
Level3 Percentage: 81% - 32v
Level4 Percentage: 100%
Soft Start Speed: Med
Memory Start: Yes
Vlow: 40v
Vlow behavior: Pulse
Vlow Timed shutdown: No
Vlow Undervolt shutdown: Yes 94%
Overtemp Warning temp: 80C
Overtemp Shutdown: No
Response: default
------------------------------------------
NEW X1 REGULATOR #2, and #3 for 4559 bulb
Vbulb(max): 29.5v
Level1 Percentage: 50% - 15v
Level2 Percentage: 69% - 20v
Level3 Percentage: 81% - 24v
Level4 Percentage: 100%
Soft Start Speed: Med
Memory Start: Yes
Vlow: 37v
Vlow behavior: Pulse
Vlow Timed shutdown: No
Vlow Undervolt shutdown: Yes 94%
Overtemp Warning temp: 80C
Overtemp Shutdown: No
Response: default

 

[JimmyM]

Jimmy,
Since I hope to be first in line for these, I will take a shot at specs for mine. If you need other info, let me know.

Got it.
But you don't need to specify Vbulb(max) or other percentages any more.
The voltages are set directly as voltage numbers out to tenths of a volt.
Also, disabling the overtemp shutdown is NOT an option. If the regulator exceeds 80C it could potentially be damaged.
In testing, the Vout was amazingly close to the programmed values. Usually < 0.01V deviation as Vbat was varied.
Mmmmm... 12-bit regulation.

Here are the settings I will use...
NEW X1 REGULATOR #1 for 64663 bulb
Level-1: 20.0V
Level-2: 27.0V
Level-3: 32.0V
Level-4: 39.0V
Level-5: 0.0V
Memory Start: YES
Soft Start: MED
Vlow: 40.0V
Vlow voltage based shutdown: YES 94%
Vlow timed shutdown: DISABLED

------------------------------------------
NEW X1 REGULATOR #2, and #3 for 4559 bulb
Level-1: 15.0V
Level-2: 20.0V
Level-3: 24.0V
Level-4: 29.5V
Level-5: 0.0V
Memory Start: YES
Soft Start: MED
Vlow: 37.0V
Vlow voltage based shutdown: YES 94%
Vlow timed shutdown: DISABLED

 

[JimmyM]

Boards have shipped and the paste stencil will ship very shortly.
I have enough FETs on hand for 6 units. The remainder are still on backorder with Digikey.
Initial testing will be done with only one FET in place, farthest away from the terminals. Then I'll use my power supply to run 3 64458 or 64623 bulbs in parallel for a 20+ amp load. Even with the 20A limit on my power supply, I should be able to get >20A RMS at the bulbs since I have a big capacitor on the power supply output and the regulator uses PWM. If one FET can do 20A, then 2 will do 40A. I can't afford to buy a big enough power supply to test these things at full power.
Once testing has concluded and I'm satisfied with the power handling, I'll post a pre-order end date.

 

[JimmyM]

Well,
Welcome back everyone. Due to the recent crash, we've lost a bunch of good posts.
But here's where things stand.
All orders have been received and shipped as necessary.
I have built a little website for keeping support files.
http://mysite.verizon.net/jminihane/id2.html You can find instructions, calibration factors, diagrams, etc there.

 

[Colonel Sanders]

Good to be back! :buddies:

Jimmy, you know I gotta ask how it's looking on doing another run of M@G regulators.

 

[JimmyM]

Good to be back! :buddies:

Jimmy, you know I gotta ask how it's looking on doing another run of M@G regulators.

Well, After these are sold (another 12), I will be researching the possibility if a C regulator. Depending on how well that goes, I'll be either building that or another run of the D1s.
I've been kicking around an idea for a D2, but that's still just an idea.

 

[Colonel Sanders]

I'd be up for a Ds or Cs. So how would a D2 be different than a D1?

 

[Starlight]

SN001 is tested and working perfectly. I am currently trying to mount the bulb socket in a Big Beam reflector. SN002 is installed in a new Larry12K. It also works perfectly. SN003 is waiting on a new host. These are VERY nice, Jimmy. Thank you.

 

[JimmyM]

SN001 is tested and working perfectly. I am currently trying to mount the bulb socket in a Big Beam reflector. SN002 is installed in a new Larry12K. It also works perfectly. SN003 is waiting on a new host. These are VERY nice, Jimmy. Thank you.

That's great, Starlight. Thanks.
After or during a run, how warm do the FETs get?

 

[Starlight]

Jimmy, I don't know. The board is buried inside the light and I have never tried to check.

 

[BVH]

OK, just a little plug for the data I collected on the Q4559x. I tested at various Voltage levels, typically, .5V increments to determine Watts and Lux. Jimmy was kind enough to take my data and make a great spreadsheet and post it on his site linked above. Hopefully, it will help you in programming your adapters for the results you want. All data is at the bulb. Use the Lux readings for relative, not absolute comparisons.

 

[petrev]

Great Info Site

:thumbsup:

Thanks Jimmy

 

[JimmyM]

Thanks for the feedback.
Starlight. I was just curious about the FET temperature. During testing with 50Vin @ 30A, the never got above 58C in a 23C basement. Just wondering what they were like with a lower load.

Re: Info Site.
Is there anything else you folks would like to see?
I'm working on putting together some info regarding driving LEDs from the Aux PWM outputs. LED color and accompanying resistor value.

Since sales have been slow lately, I've been messing with firmware.
I'll be posting a new Rev1 version (1b) that will have compile messages regarding options. So you can see if you have made any errors as well as just to confirm your selections.

What's coming in Rev2...
I'm adding code to support the Aux PWMA & PWMB outputs as low voltage and over temp warnings.
Low Voltage:
In addition to the Low Voltage warnings of Dim and Pulse, there will be a 3rd option to just shutdown the output (actually 1/4095th of full output) and light an LED (User configurable brightness).

Over Temp:
In addition to the Over Temp warning of Dim, there will be an option to shutdown the output and flash an LED on the other Aux PWM output.

So far things look good.
Right now the low voltage warning auto-recovers if battery voltage rises again. This has caused flashing in testing since the bulb load almost entirely removed. I'd like to have it latch in low mode until you press the button again or for a specified amount of time. Still working out the options of what would work best.

Over temp shuts down the output and flashes the LED as expected and is only recoverable by pressing and holding the button to shutdown the light.

I was also thinking of combining the low volt mode warning and over temp warnings into a single LED. Steady LED = Low Voltage, flashing LED = Over temp. That would leave the other PWM output for other stuff.

That leaves a question in my mind though. instead of 2 low voltage levels (warning and shutdown. Would it be better to use the LED as a "nearly low voltage" indicator. At a set voltage level the LED would light. As the pack voltage continues to fall, the existing low voltage warning then shutdown would occur?

Your thoughts?

 

[JimmyM]

Thank fully I ordered the parts for my new Bench PC yesterday. The old one is a pig. It's actually a 3.06GHz P4 laptop. New one is a self built AMD Phenom II X4 955 machine running XP with an SSD boot drive and a 500GB storage drive. March 25th can't come fast enough.

Also, I won't just be posting the latest firmware. I'll be posting each revision, so you can pick or troubleshoot, or develop, or what-have-you on whatever version you want.

 

[JimmyM]

I'm not sure how many of you will get this with the email notifications being what they are... But.

One of my customers has run into a situation where he has configured the Vlow behavior to Dim the bulb, but it pulses for a while THEN dims. From testing I have found out what I believe is going on.

His Vlow is 42V.

At 45V things work as expected, at Vlow, the output dims. as expected.

However, I bet I know what's happening. Under load your batteries sag, triggering the low voltage protection, so the output dims. As soon as it does, the batteries recover and the output returns to normal. Then the batteries sag again. Setting up an up/down voltage output. Finally, the batteries don't recover and the output just stays dimmed until the voltage sags enough to shut down the regulator.


This is a good thing actually. Good for software development. Annoying for him I'm sure...


I've already got a rev1b just about complete and ready to release. It's mostly just compile time configuration options output, but I can add a "VLOW LATCH" option. So that when it hits Vlow, it will not recover if this option is selected. The Vlow condition will stay set until the regulator is shutdown and restarted.


Does that work for everyone?

 

[wquiles]

That sounds like a good idea Jim. In my implementation, I filter the events (such as low battery) for like a 1/4 of a second or so, and if still valid, I then also latch the event - only a full hardware reset clears the flag (in my case removing at least one cell in the PhD-M6 pack). In fact I got the idea from an earlier code release from you and/or Alan

 

[JimmyM]

I have a configurable (at compile time) option to check for an undervolt or overtemp condition for 1/8, 1/4, or 1/2 second before acting on it.

The VLOW_LATCH functionality is now available in Rev 1b. It can be found on my web site for the X1 here The basecode file is completely compatible with your existing options file. All you need to edit in the basecode file in the reference to your options file.
The Rev 1b release simply adds compile time pre-processor messages and implements the VLOW_LATCH functionality.

 

[BVH]

Programming crew at work loading Vers. 1b

Latching function worked perfect!