DC-DC converter w/ isolation

Go To Last Post
19 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Greetings,

 

I've been given, for repair, a PSU for 24V theatrical DMX lighting accessories, which contains an isolating DC-DC converter.  I'm trying to find specs on this part, and it's proving to be elusive.

 

The part number on the package is M1WA0505S, and it looks like this:

 

MEGA's website doesn't seem to have any detailed specs or datasheets.  Only a catalogue:

http://www.megaelectronics.com/wp-content/uploads/2015/09/DC-DC-2013-single-pages.pdf

 

 

Those general specs seem to match what I'm seeing.  It is a 5-terminal device in a SIP package with a group of two terminals and a group of three terminals.  The two terminals are being fed with 5V, and the three terminals seem to be a dual output of -5V/0V/+5V.  All output terminals do indeed seem to be isolated from the input terminals.

 

This -5V/0V/+5V arrangement would seem to be supported by the part number (M1WA0505S) and the catalogue entry, although it isn't an exact match.  Maybe an obsoleted part?

 

In any event, I'd like to be certain of the specs for this part, because my analysis of the PSU in which it is found shows something very strange.  The isolated output powers an SN75176B, which is a 5V part.  The strange part is that it is powered by the M1WA0505S's -5/+5V lines.  The 0V line is unconnected.  That's 10V, confirmed with a DMM and an o'scope.

 

So I've got two possible conclusions:

  1. The M1WA0505S is not a -5V/+5V part and/or is damaged.
  2. The designers of the PSU should get an angry letter.

 

I can fix this either way.  If it's 1), I'll replace the converter.  If it's 2), I'll correct the wiring between the converter and the SN75176B.  I actually have a hunch it is 2).  Either way, I'll also replace the SN75176B.

 

Incidentally, this mystery was not actually related to the initial repair which the unit needed.  That repair is complete.  The mystery was uncovered during the diagnosis of the problem.  Believe it or not, the unit still works fine even with the transceiver powered by 10V!  Although the transceiver gets pretty hot, of course.  I have no idea how long it has been operating this way, but as I say I suspect it is a design flaw.

 

However, there is a remaining issue which is non-critical, but I believe it may be related to the over-volting of the SN75176B.  There is a 'DATA' indicator which lights when DMX is present.  It doesn't actually validate proper DMX framing, it merely lights when the 'R' output of the transceiver is high.  Under normal circumstances, without a DMX signal entering the unit, this LED would remain off.  Now, however, its behaviour is erratic.  I expect progressive damage to the transceiver is responsible.

 

Now the questions:

  1. Any ideas where I might find a datasheet for this DC-DC converter?
  2. Anyone used these converters from MEGA?  Any advice?  Issues?  War stories?
  3. Can a sparky genius provide a 3rd, possibly more sane, conclusion?  Or poke any holes in my analysis?

 

Thanks.

 

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The part looks like a standard part made by many manufacturers.
http://uk.farnell.com/tracopower/tma-0505s/converter-dc-dc-1w-5v-0-2a/dp/1007511

1W 200mA. The 0505D is the dual output. It has 5 pins.
Strangely enough, i've used the same part to do the same job.

I've not had that style of converter go over voltage on me. But some are more cheap n cheerful than others.

75176 on the other hand - look at them and they die! They tend to draw lots of current and get hot when they die.
The LT1785 has a better esd and input voltage spec -60V.

Last Edited: Thu. Nov 12, 2015 - 04:38 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
The part looks like a standard part made by many manufacturers.
http://uk.farnell.com/tracopower...
Marvellous.  Thanks for the link.

 

Kartman wrote:
1W 200mA. The 0505D is the dual output. It has 5 pins.
So the 'S' would be the single?  While that makes sense, two things make me question that:

 

1) The MEGA catalogue shows the M1WA_SR2 series to be dual output devices, with the single output version in the M1WB_LSR2 series.

 

2) While the TRACO power datasheet shows that, for single output devices, the centre pin (pin 5) is absent, for the dual output devices that pin is present and is the common to the -ve and +ve outputs.  What's more, I've measured -5V between pins 5/4, and +5V between pins 5/6.

 

Kartman wrote:
Strangely enough, i've used the same part to do the same job.
The TRACO or the MEGA?

 

Kartman wrote:
I've not had that style of converter go over voltage on me. But some are more cheap n cheerful than others.
I can characterise MEGA parts neither as high-quality, nor as cheap n cheerful, as I have no experience with them.  However it would seem to me odd for a single-output device to fail over voltage at exactly x2 the rated output, and to somehow grow a common.

 

Quote:
75176 on the other hand - look at them and they die! They tend to draw lots of current and get hot when they die.
So I've heard. 

 

Quote:
The LT1785 has a better esd and input voltage spec -60V.
I may consider replacing all 4 transceivers (the isolated 'in' and the three 'outs') with the Linear part, but the reality is that the theatrical environment just isn't as demanding as your typical industrial application.  I've rarely seen better than the 75176B in any of the gear I've opened up, whether a desk, a dimmer card, or a fixture.  There was a Radiance hazer which had a different transceiver.  I don't think it was a Linear part.  Whatever it was, it died an awful death for no discernible reason (possibly when it was rented out to another theatre).  I replaced it with a 75176 I had in my parts bin last year and it has been happily running ever since.

 

Many thanks for your help.  I'm still uncertain what to conclude.  A curiously over-volting converter, or a particularly tenacious 75176B which has survived a really stupid design error in an old, low-volume product.  From my experience with this kind of gear, I'm inclined to go with the latter.

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

Last Edited: Thu. Nov 12, 2015 - 05:33 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've used neither of those brands, but a few others. There a fairly genetic part.

I'd agree its probably the wrong part put in - be it wrongly marked by Mega Inc or the equipment manufacturer. Stranger things have happened before.

In the late 80's i designed a few dimmers and protocol converters that used 75176s with no protection. I had a few failures. Inthr 90's i was doing industrial controls, so i added transzorbs for protection. Even devices like the max485 die reasonably easily, so i'd recommend adding tranzorbs on the bus. I had some equipment die from lightning in China a few years ago. The max485 and tranzorbs died. Turns out the 0V wire was left out that allowed too great a voltage to be developed.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
I'd agree its probably the wrong part put in - be it wrongly marked by Mega Inc or the equipment manufacturer. Stranger things have happened before.
It may just be my own bias, but this is looking more and more likely.

 

I'm not sure why I didn't clue into it before, but the other repair I made might be related.  And a number of anomalies, in my mind, just reinforce this conclusion.

 

The muscle in the PSU is provided by what look like three off-the-shelf SMPS units, each bolted to the chassis.  They provide 24V to the main accessory power bus, via three beefy diodes, I suppose so that ripple and switching from one doesn't affect the other two, and so that a failure in one or two of them allows the unit as a whole to continue to work.

 

The 24V is regulated down to 5V through a peculiar arrangement of diodes and regulators.  First, the 24V passes through a zener, then into a 7815 with heat sink  The output of the 7815 passes through another zener, then into a 7805 with heat sink.  The 5V output drives the converter, the three output transceivers, a 6N137, and the 'DATA' LED.

 

Now, the zeners are a guess.  Those original parts aren't even on the board any more.  There is a silk-screen profile of them suggesting quite beefy through-hole parts, and the cathode marks are clearly on the high-side, suggesting they were zeners meant to dump a fair bit of heat, rather than standard diodes meant to protect against polarity reversal.

 

Those parts are gone.  It looks like a pretty inexpertly done repair, but the first zener has been replace with a low-value, high-wattage resistor in series with a what looks like a normal, small glass-package diode, forward biased, being pushed beyond it's specs.  The voltage drop across the diode is about a volt, and the drop across the resistor is about 3.5 volts.  The zener between the 7815 and the 7805 has been replaced by two small glass-package diodes in series, one of which is appears to be a zener, the other of which is likely a normal forward-biased diode.  The voltage drops are similar.  It was the mid-air solder joint between these two diodes which failed and resulted in the unit ending up on my bench.

 

Your hint that the 75176 failures result in excessive current draw makes me conclude the following:

  • Design flaw on PCB, or incorrect MEGA part (pulled from wrong bin by hand-assembler, or mislabelled by manufacturer, or shipped in error by vendor and not caught during assembly)
  • Excessive current drawn by 75176 caused eventual failure of beefy zeners
  • At least one past repair to address a frequently failing unit.  I expect the 75176 has been replaced at least once, as it (and one of the 'send' transceivers) is a SN75176BP, while the others are DS75176BTN.  I'd posit that one of the first repairs may have replaced the dead DS75176BTN with one of the send transceivers, leaving one of the three sends unusable.  A later repair, after that one also failed, would have replaced both units.

 

Who knows.  All the evidence is circumstantial, of course.  I've sent an email to MEGA in the hopes that they can give me specs on this specific part number.  For now, I'm inclined to replace all the transceivers, modify the PCB to power the isolated one from 0V/+5V instead of -5V/+5V, and possibly replace the mess of diodes and linear regulators with an off-the-shelf SMPS module to drop the 24V to 8V or so, and then use a fresh 7805 to get a clean 5V.

 

Oh, and the silk-screening on the PCB is reversed!

 

Not a single tranzorb to be found, by the way.  I'm not likely to go to the trouble of installing any.  I'm not getting paid enough.

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

Last Edited: Thu. Nov 12, 2015 - 02:35 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I use quite a few isolated DC-DC converters from XP Power. Never had a problem.

#1 This forum helps those that help themselves

#2 All grounds are not created equal

#3 How have you proved that your chip is running at xxMHz?

#4 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With regArds to your original post Joey, can't you simply call MEGA?

Jim

If you want a career with a known path - become an undertaker. Dead people don't sue! - Kartman

Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The Pic shows the LS part number, the one below the yellow highlighted part.

A quick call the maker would answer your question in less then 5 min. I would think.

 

Jim

Click Link: Get Free Stock: Retire early! PM for strategy

share.robinhood.com/jamesc3274

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ki0bk wrote:
The Pic shows the LS part number
The pic is from the intertubes for reference, not a photo of the part in front of me:

http://www.jotrin.com/technology/details/1_W_converters_target_IGBT_driving_circuits

The SIP form factor is the same, but the part number is not.

 

jgmdesign wrote:
With regArds to your original post Joey, can't you simply call MEGA?
joeymorin wrote:
I've sent an email to MEGA in the hopes that they can give me specs on this specific part number.

 

It was late when I sent them an email yesterday.  I've called today and spoke with a rep, and am now waiting for him to either call back or email me some info.  Frankly, I'm not holding my breath since it doesn't represent any new business to them (in response to my initial question about specs, the first words out of the rep's mouth were "how many units are you looking to purchase").  At best, if I conclude that their component does indeed need replacing, they'll sell me one piece.  That won't even cover the cost of time they've already spent talking to me on the phone.

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

Last Edited: Thu. Nov 12, 2015 - 08:33 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, I'll be.  I was too hasty.  The rep sent me an email this afternoon, with the datasheet attached.

 

As I suspected, this is indeed a dual output device:

 

 

 

 

The conclusion seems unavoidable.  Design flaw, or improperly shipped/picked component during assembly.  I'll rejigger the PCB with some green wire, replace the transceivers, and scrape the mess of zeners, resistors, and regulators off with a spatula so I can replace them with a switcher and possibly a linear reg to get to a clean 5V.

 

Thanks for all your feedback.

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Good to hear Joey. One thing to be aware of is the current. With two 120 ohm terminators, you'll be running close to 100mA on the 5V. This means you'd be running the converter close to max. In reality this probably wont be an issue, but i thought i'd mention it.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks for pointing that out.  Actually, the converter only powers one of the SN75176B transceivers, and nothing else.  That transceiver is permanently rigged for receive, so it will never drive the bus.  DMX is a single-TX-node protocol.  Only the transceiver in the lighting desk (or an inline opto-spiltter appliance) drives the bus. RDM is a follow-on which allows multi-node TX, but under co-ordination of the lighting desk master.  This accessory PSU doesn't support RDM, and none of the accessories it is meant to drive do, either.

 

The only other load is the input LED to a 6N137 opto, driven by the isolated 75176B's RO pin.  The LED is in series with a 360R resistor, which with a Vf of 1.4V and a Vcc of 5V should result in a forward current of about 14 mA.  However, driving at 10V, the current has been more like 24 mA, which is beyond the absolute maximum for the 6N137.  I might just replace it, too.

 

Total load on the converter >>should<< be about 14 mA for the 6N137 input, plus a max of 35 mA for the 75176B, or about 50 mA.  Likely the load has been much higher, since as you point out a failing 75176 draws much more current, and that is likely exacerbated by 10V!

 

What I will need to be careful of is ensuring the converter has a >>minimum<< load of 10 mA on each leg (specified in the datasheet).  I'll just add a 470R resistor on the -ve and +ve legs to ensure this, even if the 75176B dies open circuit in the future.

 

During testing yesterday I was running it with no load for several minutes (less than an hour), before I knew that I shouldn't.  I hope that hasn't damaged it, but since it has likely been operating for years at well beyond rated current, I expect it is fine.

 

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I believe that you should assume that the designer didn't know what they were doing and happened upon a combination of parts and settings that solved the problem at the time, and went with that design through the life of the product.  This is not that unusual in small companies with limited resources, especially before the internet.

 

I suggest rebuilding the circuit to specification with new parts.  Then confirm that it is working to customer's expectations.  Adjust the design and replace the parts on everything that is running out of spec.   Contact megaelectronics or other companies that make small isolation DC-DC converters for test procedures.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Replace this part...

 

Replace this part...

 

Replace this part...

 

It seems you have found a flaw in the original design, and that it was a slow be materialize but catastrophic design failure.

 

You also have an excellent knowledge of what the PCB is suppose to do, and the protocol driving it.

 

At some point would it not be easier to just redesign the PCB to replace the original PCB?

I'm thinking same mechanical design to fit in the box, drive the same outputs, etc, but use your circuitry.

 

Perhaps there is a market to retrofit the older units?

Perhaps you can purchase a bunch of failed older units and refurb them with your new PCB?

 

Perhaps you will just sleep better knowing that the newly refurb'ed units are not going to fail in the middle of a show.

 

JC

 

Edit: Typo or two

Last Edited: Fri. Nov 13, 2015 - 05:08 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Running those converters without a load normally causes the output to rise a little bit as it doesn't regulate. Looks like you're getting closer to DMX nirvana.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Simonetta wrote:
I believe that you should assume that the designer didn't know what they were doing and happened upon a combination of parts and settings that solved the problem at the time, and went with that design through the life of the product.  This is not that unusual in small companies with limited resources, especially before the internet.
I expect you're not >>too<< far from the truth.  As a theatre tech, I've probably handled hundreds of these units.  They rarely fail.  In all honesty, I generally have greater faith in the older units than the newer ones.  I've seen a few new units die the very week they were pressed into duty.

 

This particular unit is, I expect, one of the very first produced, with the kinks worked out of the later revisions after too many from this one made it back for repair.  There's no way for me to know what it's service history and repair history are.  Some of the evidence of repair I've seen could have been done by the manufacturer, or by the vendor, or by the rental house from which this unit was later purchased second hand, or by an in-house technician.  I'll never know.

 

Quote:
I suggest rebuilding the circuit to specification with new parts.
Were it my own, that's probably what I'd do.  As it is, I'm not getting paid enough for the time I've already spent on it, and will soon spend to do the planned repair.

 

DocJC wrote:
At some point would it not be easier to just redesign the PCB to replace the original PCB?

I'm thinking same mechanical design to fit in the box, drive the same outputs, etc, but use your circuitry.

 

Perhaps there is a market to retrofit the older units?

Perhaps you can purchase a bunch of failed older units and refurb them with your new PCB?

The thought has crossed my mind.  I'll talk to the TD who hired me to fix this one to see if he knows of a pool of failed gear I could dive into.

 

Quote:
Perhaps you will just sleep better knowing that the newly refurb'ed units are not going to fail in the middle of a show.
I certainly would.  For now, I'll have to be satisfied that I've returned it in better condition that it was when received it, and almost certainly in better condition than it was when it was first switched on.

 

The notion of trying to supplement my income by fixing/upgrading this kind of gear is on some level appealing, but probably about as lucrative as a career in VCR repair ;-)

 

I've got a couple of Rosco I-Cue mirrors, and a couple of DMX iris units in a 'when I get a round tuit' pile.  I'm sure I'll have questions when I get to them, but it's been over a year so far, so nobody hold your breath!

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

Running those converters without a load normally causes the output to rise a little bit as it doesn't regulate. Looks like you're getting closer to DMX nirvana.

I did notice the rise, but what concerned me was later reading this:

Requirement on output load
    To ensure this module can operate efficiently and reliably, During operation, the minimum
output load is not less than 10% of the full load, and that this product should never be
operated under no load!
If the actual output power is very small, please connect a resistor
with proper resistance at the output end in parallel to increase the load, or use our company’s
products with a lower rated output power (MW2A_S /MW2B_LS series).

Overload Protection
    Under normal operating conditions, the output circuit of these products has no protection
against overload. The simplest method is to connect a self-recovery fuse in series at the
input end or add a circuit breaker to the circuit.

Their emphasis.

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The manufacturer has spoken!
I do a couple of repairs now and then - especially if it is something interesting like analog synths or old video games. I'm definitely not going to get rich doing it!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I bet the original design specified the single output device (M1WB0505LS), but someone down the line noticed it isn't UL certified and entered a change request, the impact of which was not properly analysed.  The addition of a 470R resistor and the moving of a single trace would have solved the problem.

 

"Experience is what enables you to recognise a mistake the second time you make it."

"Good judgement comes from experience.  Experience comes from bad judgement."

"Wisdom is always wont to arrive late, and to be a little approximate on first possession."

"When you hear hoofbeats, think horses, not unicorns."

"Fast.  Cheap.  Good.  Pick two."

"We see a lot of arses on handlebars around here." - [J Ekdahl]