That datasheet is for the high-frequency (30-60kHz) version, this is the one I have.
Thanks
-
Steel Soldiers now has a few new forums, read more about it at: New Munitions Forums!
MEP-831a Inverter Overhaul
- Thread starter marcjs14
- Start date
Thanks
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Thank you Marcel,
That’s a good idea with the H Bridges. Not to rain on your parade, but Ceramic Capacitors, which are quite different compared to electrolytic capacitors, have usually a MTBF (Mean Time Between Failure) of 100,000 plus hours if they are of decent quality.
Security Capacitors for power applications have a required MTBF of 1,000,000 hours.
So 1,300 hours is pitiful at best.
Have a great evening
Peter
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
Thanks, yes I saw that, but I posted the new version of this part which is pin compatible with the electrical values pretty much the same.
I am glad you were able to find the part NOS
Peter
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Since I called this inverter design “piss poor” in post #53 after Kurt’s report about the non existing inrush current limiter,
I figured I am going to publish a properly designed powers supply for such an inverter but without any additional EMI filter for simplicity reason
the safety/security capacitors are basically fail safe ceramic capacitors which prevent exactly this of what Marcel and possibly other owners have experienced. A safety capacitor is on average 20 to 30 times more expensive then a standard ceramic disk capacitor especially when they have agency approval like UL and/or are CE compliant
i don’t think I have to explain anything beyond this.
I figured I am going to publish a properly designed powers supply for such an inverter but without any additional EMI filter for simplicity reason
the safety/security capacitors are basically fail safe ceramic capacitors which prevent exactly this of what Marcel and possibly other owners have experienced. A safety capacitor is on average 20 to 30 times more expensive then a standard ceramic disk capacitor especially when they have agency approval like UL and/or are CE compliant
i don’t think I have to explain anything beyond this.
Last edited:
Peter, I like it although I wouldn't know where to start to implement it!
Sticking to the basics, I've sourced a sh*tload of Vishay S103K75Y5PN65K0R caps to replace the Hitano garbage as I pick through the rebuild. I also ordered replacement Vishay 600V 15A Hyperfast FREDs just in case the originals were damaged..
If after all this, the thing goes up in smoke again...at least the baby carp in Lake Norman will have some shiny objects to look at in their underwater museum.
Sticking to the basics, I've sourced a sh*tload of Vishay S103K75Y5PN65K0R caps to replace the Hitano garbage as I pick through the rebuild. I also ordered replacement Vishay 600V 15A Hyperfast FREDs just in case the originals were damaged..
If after all this, the thing goes up in smoke again...at least the baby carp in Lake Norman will have some shiny objects to look at in their underwater museum.
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
I think it will be fine your rebuild. As long as we don’t miss any remaining defective parts or blown traces.
The Vishay’s are an excellent choice and comply with:
IEC 60384-8 IEC 60384-9
Which will assure they are very high quality caps.
On the hyperfast freds you can also go with a higher current rating then 15 Amps. It may cost a few bucks more for the whole batch but the higher the current rating the less stress the diode experiences during communication
Peter
I will see whether I can change the FRED order to the 30A version (VS-E5PX3006L-N3). They are a bit slower, but not by much.
Thanks!
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
The lake of resin is conformal coating. Most likely a acrylic resin based conformal coating.
This stuff here works quite well on getting the conformal coating removed:
Conformal Coating Stripper
this will make everything much easier when you strip all board completely
you can then apply new conformal coatings when everything is done and properly working
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
The Fred’s and similar devices were designed for fast switching power supplies and the 60Hz they have to contend with here is more like molasses in comparison.
THE STRIPPER IS ON THE WAY!
I thought I'd make my morning post sound like fun...speaking of which - I'd had more fun in H.S. detention!
Those electrolytics were a PITA to pull. I think I'll wait for the stripper to arrive before I attack the diodes and ceramic caps...I am still trying to catch the order to upgrade the diodes to 30A before they ship.
That's all for today.
I thought I'd make my morning post sound like fun...speaking of which - I'd had more fun in H.S. detention!
Those electrolytics were a PITA to pull. I think I'll wait for the stripper to arrive before I attack the diodes and ceramic caps...I am still trying to catch the order to upgrade the diodes to 30A before they ship.
That's all for today.
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
great job! When you strip the board, check the 4 resistors. The color strips are not clearly visible in the picture, but they look like green, brown, orange and gold. Gold stripe means they have 5% tolerance. Once you confirm colors then I can tell you what value they have. They carbon resistors, that I can tell you already.
The burnt / charred area of board can be carefully cleaned off with MEK or Flux Remover for Rosin based flux. You can then fill the cleaned up area with just regular fiberglass epoxy resin. This way the board will be level again and the exposed fiber mesh in the middle will be protected.
Who did you order the Diodes from? Newark?
Peter
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
the other thing which would of interest are the twelve e-caps. Do you have a ESR tester? If not just regular capacitance test would be fine and see what that yields. Thanks
Peter
the other thing which would of interest are the twelve e-caps. Do you have a ESR tester? If not just regular capacitance test would be fine and see what that yields. Thanks
Peter
Thanks Peter!
I think I'd make the resistor bands to be blue, green, orange, and gold. Green could be brown, but I doubt it...In any event, I'll verify once pulled from circuit mid-week when other parts arrive,
I've ordered the caps and diodes from Arrow Electronics, which is my primary source for electrical components. I believe them to be reputable (despite them being a "global" corp).
Peter, no ESR tester - just a crappy Chinese knock-off capacitor tester with a yellow case made to look like a Fluke...
...assuming the tester tells the complete story, no smoking gun here as the caps all test well within their +/- 20% spec (worst-case within 3%)...
Thanks,
Marcel
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
Thanks, just as I figured. These are 560uF caps. The values your tester came up with are most likely correct.
And yes, they have a 20% tolerance, but all those caps are basically toast.
Let me explain:
The dry Electrolytic Capacitor in its construction and make up was invented around 1925 by Ruben and Mallory. Commercial production started in 1930. Almost a 100 years ago.
Anyway, the construction was the same, aluminum or tin can with a fabric soaked in electrolyte layered in between paper and mica strips.
These e-caps were hand made by mostly woman back then. Depending on how long the Ladies soaked the fabric in the electrolyte and the actual thickness of the mica sheets, the end capacitance after assembly varied greatly from capacitor to capacitor of the same type.
The manufacturer, in the US it was Cornell Dublier and in Germany AEG, specified the acceptable tolerance between +70% and -50%
So, the caps you got, which are basically toast would have been ultra premium quality.
With the introduction of automation in manufacturing after WW II, the tolerances decreased to 20% in the 1960’s
That was 62 years ago.
The capacitor manufacturers left it by that since, mostly for convenience reasons, meaning their bank accounts.
Now, with the automation technology these manufacturers have employed since the 1990’s and the introduction with LEAN, the tolerances are 1% even though they print 20% on them.
This tolerance applies only to newly manufactured electrolytic capacitors and is only guaranteed for app. 1000 hours of operation where storing or not operating a device counts toward these hours at a de-rated value depending on actual temperature during the storing / non operation.
as a general rule figure 2 years of operation and 5 years of storage in a climatized warehouse.
So, these are 560uF and basically if they are at 566uF or 554uF then they are toast.
yours are very toast.
The reason for this deteriorating is the introduction of the safety vent in electrolytic capacitors in the early 1980’s.
The caps are no longer fully sealed.
Peter
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
the resistor would be then 65 KOhm 2W which are most likely bleed resistors to drain the DC Bus from around 300V to OV when the unit is shut down.
The one in the picture is chipped. You may want look into replacing them too.
KOA or Vishay would be recommended.
Will be interesting to see what their actual resistance is when you desolder them
Silly me! When I started measuring them and they all came in around 500 I completely forgot they were supposed to be 560uF and not 500. To be consistently off by on average 12% is in itself interesting. Maybe a dry 560 becomes a 500 - in which case they even fail the same way???
I will measure the resistors as I pull them and will replace with a suitable quality part.
thanks,
Marcel
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Marcel,
it depends on how they dry out and where the safety relieve is located in relation to the terminals. That’s a science in itself.
What is upper capacitance value your tester can test to?
The range goes up to 20mF.
I've been scouring the internet for 65k 2w ceramic resistors, and they appear to be unicorns. How critical would the value be (62k and 68k readily available)?
- 1,118
- 1,244
- 113
- Location
- Basehor, KS
Is this 20 mili (e-3) Farad or 20 mikro (e-6) Farad? The Chinese are not used to our numerical system.
As to resistor, use this one here:
A few Kilo ohm up won’t matter, more important they are rated for 500V
We will see what you actually got when you take the old ones out, your green may turn out to be gray and that would 68 KOhm to begin with.
20 mili...it goes from 2000 pF-->2 uF-->20 mF in ten steps.
Done it. I wasn't able to change my order to the 30A FREDs since Arrow mysteriously sold out overnight, so I ordered them thru mouser along with the metal film resistors. I'll just return the 15A FREDs after they arrive.
Thanks,
Marcel