HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Mar 07, 2014, 02:30 PM
Registered User
Seattle
Joined Sep 2007
2,178 Posts
I would also check the Meanwell under heavy load. The load from charger is not constant. The charger pauses (very briefly) about 20 times per second to take readings. During the pause, the load on the supply is lifted. If the supply's regulator is not fast enough, the supply voltage can easily over shoot the target voltage. Usually the overshoot and recovery occurs far to quickly to see it on a DVM, but you should be able to see it on the scope.
Gregor99 is offline Find More Posts by Gregor99
Reply With Quote
Sign up now
to remove ads between posts
Old Mar 07, 2014, 03:07 PM
Registered User
Joined Dec 2009
189 Posts
i'm not familiar to this subject so i need to ask, please:
1) is the ideal situation never to have any sort of spike when turning on a ps?
2) voltage should only differ from zero when the charger is actually charging a battery (i.g., when there is some current)?
thanks.
Can'tFly is offline Find More Posts by Can'tFly
Reply With Quote
Old Mar 07, 2014, 03:42 PM
Registered User
USA
Joined Jan 2002
4,700 Posts
Quote:
Originally Posted by Gregor99 View Post
I would also check the Meanwell under heavy load. The load from charger is not constant. The charger pauses (very briefly) about 20 times per second to take readings. During the pause, the load on the supply is lifted. If the supply's regulator is not fast enough, the supply voltage can easily over shoot the target voltage. Usually the overshoot and recovery occurs far to quickly to see it on a DVM, but you should be able to see it on the scope.
Aren't there also some input caps on the charger that would present an initial load on the PS when it's powered up with the charger connected or when the charger is attached to the power supply.
Ohmic is online now Find More Posts by Ohmic
Reply With Quote
Old Mar 07, 2014, 05:52 PM
Registered User
Seattle
Joined Sep 2007
2,178 Posts
Quote:
Originally Posted by Ohmic View Post
Aren't there also some input caps on the charger that would present an initial load on the PS when it's powered up with the charger connected or when the charger is attached to the power supply.
Good point. Yes, there are capacitors in the charger which will fill rapidly when connected to a live power source. This won't damage the charger and in fact was one of the scenarios in my test scenarios using a 48v source.

But as it relates to the overshoot, the charging of the capacitors could cause the supply to overshoot. I believe Michael already reported some overshoot on the Protek when the second supply was connected.
Gregor99 is offline Find More Posts by Gregor99
Reply With Quote
Old Mar 07, 2014, 09:05 PM
Michael
United States, ME, Wells
Joined May 2008
1,075 Posts
Light to No Loads, So Far

As mentioned a few posts ago, except for a quick verification that I could charge a 3S2200 LiPo, I've only been powering up/down to change presets (order, default values, etc.)

All my problems have been just connecting DC to the right half after (second unit) disconnecting DC power before unplugging the AC input. I doubt that the Meanwell, with virtually no load, would go above 75v as Howard indicated seems to have happened. I just do not believe that is possible. I also do not believe that the much lower voltage Protek 40 (nominal 13.8v) PS could, somehow, exceed 75v.

I must admit to being uncomfortable connecting the DPL8 halves to the 48v supply. The spark upon connection is disconcerting; similar to a 5S or 6S ESC. For my models, I usually use Esprit Model Anti-Spark 4mm or 5.5mm bullet connectors. I do understand capacitors on the DC input side, but I would prefer to let the PS handle that by connecting the DC leads, with AC off, then connecting AC; as I did the first round.

I have to be missing something because no scenario that anyone suggests appears able to account for the failures.

I anxiously await Howard and/or Nathan's diagnosis. In the meanwhile, I have a useless Meanwell 48v supply; probably should have gone with a 24v supply at 2000 watts; then I could have handled both the original PL8s and the DPL8. As it is, my case is a nice work of art, but totally useless.

Michael
Quote:
Originally Posted by Gregor99 View Post
I would also check the Meanwell under heavy load. The load from charger is not constant. The charger pauses (very briefly) about 20 times per second to take readings. During the pause, the load on the supply is lifted. If the supply's regulator is not fast enough, the supply voltage can easily over shoot the target voltage. Usually the overshoot and recovery occurs far to quickly to see it on a DVM, but you should be able to see it on the scope.
Snowflake6515 is offline Find More Posts by Snowflake6515
Reply With Quote
Old Mar 09, 2014, 09:42 PM
Michael
United States, ME, Wells
Joined May 2008
1,075 Posts
Protek 40 Update

Swamped, but I did connect my Protek 40 PA (13.8v) to my older PL8 units. I cycled the power on/off, with/without DC connected/disconnected, etc.; no loads, just power up / power down as I had been doing with my DPL8 units.

Zero problems. I just gave up after a while.

Now, if the lower input voltage PL8 can handle this Protek 40 supply, why cannot the higher input voltage DPL8 handle the same PS?

This is what led me to raise the point about having the red/black wires always connected between the two halves, but Gregor99 demonstrated that was not an issue.

Greg, based on your experiment with expansion servo lead, I went back to the Protek 40 on my PL8, as reported above.

That leaves only the Meanwell 48v supply; i.e., perhaps the fried component is happening upon disconnecting the DPL8 from the 48v supply and not when connecting to the Protek 13.8v supply. For the second failure, I connected / disconnected the DC side of the DPL8 while the AC had been plugged in and the DC stabilized.

Dave, at Progressive RC, has many of the Meanwell 48v 2000w PS in the field with DPL8 units and no reported problems. So, either I have a bad Meanwell 48 or something is still missing with the DPL8 connection procedure.

I should have time to 'scope the Meanwell tomorrow, Monday.

In the meantime, I am hoping that Nathan and Howard can come up with a definitive scenario that is causing the problem. It is starting to get warmer so flying season may be upon us and I would really like the DLP8 to be the heart of my battery management system.

Michael
Snowflake6515 is offline Find More Posts by Snowflake6515
Reply With Quote
Old Mar 11, 2014, 12:45 AM
Registered User
Georgia
Joined Jun 2007
649 Posts
Quote:
Originally Posted by Snowflake6515 View Post
As mentioned a few posts ago, except for a quick verification that I could charge a 3S2200 LiPo, I've only been powering up/down to change presets (order, default values, etc.)

All my problems have been just connecting DC to the right half after (second unit) disconnecting DC power before unplugging the AC input. I doubt that the Meanwell, with virtually no load, would go above 75v as Howard indicated seems to have happened. I just do not believe that is possible. I also do not believe that the much lower voltage Protek 40 (nominal 13.8v) PS could, somehow, exceed 75v.

I must admit to being uncomfortable connecting the DPL8 halves to the 48v supply. The spark upon connection is disconcerting; similar to a 5S or 6S ESC. For my models, I usually use Esprit Model Anti-Spark 4mm or 5.5mm bullet connectors. I do understand capacitors on the DC input side, but I would prefer to let the PS handle that by connecting the DC leads, with AC off, then connecting AC; as I did the first round.

I have to be missing something because no scenario that anyone suggests appears able to account for the failures.

I anxiously await Howard and/or Nathan's diagnosis. In the meanwhile, I have a useless Meanwell 48v supply; probably should have gone with a 24v supply at 2000 watts; then I could have handled both the original PL8s and the DPL8. As it is, my case is a nice work of art, but totally useless.

Michael
"Meanwell, with virtually no load, would go above 75v as Howard indicated seems to have happened."
That would indicate a "broken" supply, or a poorly designed one. The maximum OVP voltage is (spec) ~67v.
Since that is obviously over the charger's max input voltage, it looks like the Meanwell needs an external OVP circuit to
prevent significant OV. The innards of the supply should not permit an over voltage condition of that order (opinion).
If nothing else the capacity of the filters in both the power supply and the charger should not respond fast enough to allow such an over voltage before protection circuits kick in. It may also be that a minimum load is needed, or a bleed resistor in the supply is open or the wrong value.

My TDK Lambda supply would go off line if a smaller voltage overshoot occurred. And stay off line until it was reset, or power cycled.
But, it's a 24v supply. I checked it's specs, and with two in series, if either one went above 30-34v both would shut down. So that's not much different.
Two of the Lambda's can be used in series, since the supply outputs can "float", or either polarity be tied to ground.
When this is done, the supplies sense and control circuitry has to be setup (Connected and jumpered) properly, to insure synchronized operation.

Back three decades or so ago, I was part of an engineering avionics development team
We had a fairly unique power supply requirement - - 400hz 3phase 230vac in, and a very highly regulated and controlled
output of ~13KvDC at least one amp, and up to about 100A short term peak. The supply had to shut down in milliseconds, and tolerate an almost direct short on the output. It had a designed in capability to "clear the short" under a set of specific conditions.
If this was note enough , design operating altitude was in excess of 40,000 feet.
When all was said and done, the supply was encapsulated (altitude and HV don't get along) and less than 1/2 cu ft.
If the supply regulation utterly failed, resulting in a probable OV condition, a crowbar circuit clamped the output voltage
to a safe level long enough to get the supply to shut down.
Even with all of that, we did have a few catastrophic failures that actually melted holes in the aluminum case of the "black box" that the power supply
was part of. When a power supply and it's load have costs exceeding ~20k each, cascade failure is not desired!
chuck75 is offline Find More Posts by chuck75
Last edited by chuck75; Mar 11, 2014 at 01:40 AM.
Reply With Quote
Old Mar 11, 2014, 02:12 PM
Michael
United States, ME, Wells
Joined May 2008
1,075 Posts
No Load Meanwell 48v 2000w PS On 'Scope

I put the Meanwell 48v PS on my 'scope. The worst case, peak-to-peak variation was 2.4v and that was inserting/removing/re-inserting the AC cord. The 'scope stores Vmax, Vpeak-to-peak, etc.

When just powering it up and letting it sit, the typical maximum variation is about 380mv; seems to be about the spec in the manual.

I did adjust the output voltage to the minimum, 40.6 volts, to give me some head room when I get my first DPL8 (repaired) back. Previously, the default output was measured at 47.9v.

I do not have a suitable load available, but will grab some auto taillight bulbs and fabricate something suitable (similar to how I drain LiPos prior to discarding.) At that time, I will repeat the measurements as Gregor99 suggested a few posts back.

However, the tests, today, simulate the conditions associated with my failures; the DPL8 units, with no load, draw very little current.

I have removed all the USB/Expansion channel wiring as Howard was concerned that the open leads could be sensitive to static electricity. With the "Y", switch for expansion mode (did work great, going to miss that), and bulkhead access for FUIM3/USB access, there probably was about two feet of unterminated cable when USB interface was not connected.

One item I did notice, when reading the Meanwell 48v manual, again, was that the over-voltage protection is listed as: 57.6 to 67.2 volts, and will shutdown after 5 seconds. That seems quite high. However, at no time did my Meanwell shutdown requiring a restart. Also, although no load applied, the maximum peak-to-peak voltage I saw stored in the 'scope's table was 2.4v; again, should not be a factor.

Thanks for all the suggestions, folks. It has given me some areas to check, but unfortunately, nothing is popping up that would explain exceeding 75v as Howard indicated had to have happened with the first DPL8 failure.

Greg, I believe you indicated that you were using a 48v PS with no problems. If not a Meanwell, what is the over-voltage spec? Is it possible that a very short over-voltage, lasting less than five seconds, might have occurred, and a defective PS went beyond the 67.2v spec? Not sure how or under what circumstances that could occur, but if I could figure out to reproduce that, then my 'scope should store the value for later access.

Michael
Snowflake6515 is offline Find More Posts by Snowflake6515
Reply With Quote
Old Mar 11, 2014, 02:27 PM
Registered User
USA
Joined Jan 2002
4,700 Posts
Is your Mean Well running off of 120V or 240V AC?
Ohmic is online now Find More Posts by Ohmic
Reply With Quote
Old Mar 11, 2014, 02:30 PM
Registered User
Seattle
Joined Sep 2007
2,178 Posts
I'm running two DPS-1001AB supplies in parallel. The specs give the OVP at 51-55v. But don't the duration before the OVP is triggered.

Also remember that the Powerlabs pause briefly during the charge cycle (about 20 times per second). The brief lift of the current can result in a very fast (and short) overshoot from the supply. You may be able to see this on the scope. The overshoot will vary based on load and the abilities of the supply's voltage regulator.
Gregor99 is offline Find More Posts by Gregor99
Reply With Quote
Old Mar 11, 2014, 05:23 PM
Michael
United States, ME, Wells
Joined May 2008
1,075 Posts
120v ac

120v AC, so limited to about 1700 watts, total; given the 15amp circuit...

I have a separate PS for 220 that can be used in the workshop, but I find that I rarely need more than 1200 - 1500 watts, total, at any given time as I tend to charge at 1C. My larger birds use smaller 4S to 6S packs in series.

I will be purchasing a Honda EU2000i generator for field use; about 1600 watts, so the Meanwell and DPL8 will be loafing.

Michael
Quote:
Originally Posted by Ohmic View Post
Is your Mean Well running off of 120V or 240V AC?
Snowflake6515 is offline Find More Posts by Snowflake6515
Reply With Quote
Old Mar 11, 2014, 05:25 PM
Michael
United States, ME, Wells
Joined May 2008
1,075 Posts
Will Measure / 'Scope...

Thanks for the heads-up, Greg. I will put the DPL8 on my Meanwell PS and repeat the 'scope tests while also adding some typical loads. My simulated load would be constant and not take into account the brief pause cycles; I had forgotten about that.

Michael
Quote:
Originally Posted by Gregor99 View Post
I'm running two DPS-1001AB supplies in parallel. The specs give the OVP at 51-55v. But don't the duration before the OVP is triggered.

Also remember that the Powerlabs pause briefly during the charge cycle (about 20 times per second). The brief lift of the current can result in a very fast (and short) overshoot from the supply. You may be able to see this on the scope. The overshoot will vary based on load and the abilities of the supply's voltage regulator.
Snowflake6515 is offline Find More Posts by Snowflake6515
Reply With Quote
Old Mar 11, 2014, 07:35 PM
Registered User
Georgia
Joined Jun 2007
649 Posts
The FMIU3 charger/USB adapter is opto isolated, according to the on line listing. Based upon that, I ended up getting one for each half of the DPL8, rather than using a "Y" adapter.

Another tidbit concerning DC power supplies.
In general they do not like to sink current, only supply it.
To do both, a series and shunt regulator must be part of the supply, and configured by design
to be able to both source and sink current.
A few decades ago, I spent some time in GE's engineering, calibration, and QC labs.
We used calibrated, and repaired virtually everything in the HP instrument catalog, as well
as in house unique equipment. DC power supplies were one of the things that I became intimately
familiar with. Failure due to operator error in device production testing was the most common cause of trouble.

It's possible, under limited sets of conditions, to have a circumstance
that causes this to occur, such as accidentally setting the charger to regenerative
discharge into a DC supply. The DPL8 "normally" will come up with a message saying that it is switching to internal discharge,
depending, I suppose, on the voltage limit settings.
chuck75 is offline Find More Posts by chuck75
Last edited by chuck75; Mar 11, 2014 at 07:50 PM.
Reply With Quote
Old Mar 11, 2014, 09:12 PM
Registered User
USA
Joined Jan 2002
4,700 Posts
Quote:
Originally Posted by Snowflake6515 View Post
120v AC, so limited to about 1700 watts, total; given the 15amp circuit...

I have a separate PS for 220 that can be used in the workshop, but I find that I rarely need more than 1200 - 1500 watts, total, at any given time as I tend to charge at 1C. My larger birds use smaller 4S to 6S packs in series.

I will be purchasing a Honda EU2000i generator for field use; about 1600 watts, so the Meanwell and DPL8 will be loafing.

Michael
Sounds good, just wanted to make sure you were derating for 120V AC accordingly.
Ohmic is online now Find More Posts by Ohmic
Reply With Quote
Old Mar 11, 2014, 09:36 PM
Michael
United States, ME, Wells
Joined May 2008
1,075 Posts
Opto Isolated, Regenerative Discharge, etc.

The problem with the FUIM3 interface isn't with the adapter, but that the signal leads in the back of the DPL8 halves, apparently, go directly into their respective microcontroller. Any static electricity on that lead can fry it (that's apparently what happened to my second DPL8.) I had plugged in the wires while the DPL8 was unpowered. The switch for expansion mode was recycled and the pins crimped elsewhere, installed, with power off, etc.; so no voltages present, no soldering, but plugging in unpowered cables ("Y").

The problem, apparently, was the couple of feet of unterminated cable that could pick upsomething; I haven't a clue where as the unit was sitting on my test bench.

As for regenerative discharge, the beauty of the [D]PL8 is that regenerative discharge cannot be selected if the source is a DC PS; only if a battery is selected, then there is a regenerative discharge option. Also, my failures were without load; simply powering up/down with two different PS.

I did want to have a single bulkhead connector for the external Pb supply without having to disconnect the internal DC PS. To prevent the PS from having to "sink current", I installed two Schottky diodes in each leg of the DC PS to DPL8 EC5 connectors. I removed all that after the first failure when someone suggested not connecting the DPL8 without having stable DC ready; i.e., no on/off of AC.

Now, connecting 48v DC to the input capacitors of the DPL8 results in a healthy spark. I would prefer to have the PS handle that as, ostensibly, there is circuity built-in to handle that... Others are apparently using the Meanwell 48 PS with DPL8s and are turning the AC on/off without problem. I also believe that Gregor99 also indicated that he has done that; provided certain parameters of the PS are present...

Michael
Quote:
Originally Posted by chuck75 View Post
The FMIU3 charger/USB adapter is opto isolated, according to the on line listing. Based upon that, I ended up getting one for each half of the DPL8, rather than using a "Y" adapter.

Another tidbit concerning DC power supplies.
In general they do not like to sink current, only supply it.
To do both, a series and shunt regulator must be part of the supply, and configured by design
to be able to both source and sink current.
A few decades ago, I spent some time in GE's engineering, calibration, and QC labs.
We used calibrated, and repaired virtually everything in the HP instrument catalog, as well
as in house unique equipment. DC power supplies were one of the things that I became intimately
familiar with. Failure due to operator error in device production testing was the most common cause of trouble.

It's possible, under limited sets of conditions, to have a circumstance
that causes this to occur, such as accidentally setting the charger to regenerative
discharge into a DC supply. The DPL8 "normally" will come up with a message saying that it is switching to internal discharge,
depending, I suppose, on the voltage limit settings.
Snowflake6515 is offline Find More Posts by Snowflake6515
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Sold Brand new Work of Art Dexx micro quad flohio117 Aircraft - Electric - Multirotor (FS/W) 9 Sep 15, 2014 08:42 AM
Sold Brand New Hong Nor 6006 Servo multiflyer2 Cars - Accessories (FS/W) 0 Jan 06, 2014 07:06 PM
Help! Brand New Tarot T-2D Won't Work :( JoakimT Aerial Photography 6 Dec 18, 2013 09:32 AM
Sold **BRAND NEW**-Nelson 50 sport engine with Jett style muffler!!! A true work of art!! Mike Fortune Aircraft - Fuel - Engines and Accessories (FS/W) 0 Oct 12, 2013 10:07 PM
Sold **BRAND NEW**-Nelson 50 sport engine with Jett style muffler!!! A true work of art!! Mike Fortune Aircraft - Fuel - Engines and Accessories (FS/W) 0 Sep 24, 2013 07:41 PM