Author Topic: Vintage diode identification  (Read 1251 times)

0 Members and 1 Guest are viewing this topic.

Offline Vilnis_Radio

  • Contributor
  • Posts: 6
  • Country: lv
Vintage diode identification
« on: August 10, 2022, 11:18:24 am »
Hi there!

I have the old welding machine from 80ies I believe. It has one control circuit and it is supplied by 48AC. Suddenly it lost some functions and I inspected this circuit visually and found bad looking polarized 220mf capacitors and one completely dead diode and bunch of diodes with broken ceramics body. I searched through some welding forums and there was suggestion to change these old diodes to UF4007. I did and that does not helped.

Now I came to conclusion that the problem is absence of correct replacement diode and UF4007 is not the way to go. Circuit contains bunch of these diodes looking to be identical. So I put the rectifier instead of diode bridge to get spare diodes. When I changed UF4007 to original ones circuit starts to work as it should. The problem is that I need more of those diodes to completely change the bad looking ones.

UF4007's behave strange. In some point of circuit they seems preventing forward current. In some other point of circuit it leads to increased current and overheating cap and resistor. In the diagram D16 was one that overheated R1 and C2. D6 was blocking current to relay U5. When I replaced back to original ones both issues disappeared.

Multimeter diode test shows 0.545V. Power supply bench at constant current 0.001A shows that it opens at 0.67V.

Please help me with correct replacement choice!

I am attaching picture of these diodes and the circuit diagram.

Thanks in advance!

Uldis
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 11900
  • Country: de
Re: Vintage diode identification
« Reply #1 on: August 10, 2022, 12:50:10 pm »
Identifying the color coded diodes is somewhat tricky and not always possible / unique.

Judging from the case size, the diodes may withstand slightly more than 1 A. So the UF4007 may be too weak. The faster UF400x comes with more leakage current than the original 1N400x - this may be a problem in some cases.  With just mains frequency rectification for the electronics / relay supply there is no need for fast diode.

There seem to be still a few of the diodes that are looking good. To see if the diode is a fast one (like the UF4xxx) or a normal slow one, one could do a quick test with square wave generator and scope.
 
The following users thanked this post: Vilnis_Radio

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2972
  • Country: de
Re: Vintage diode identification
« Reply #2 on: August 10, 2022, 01:57:12 pm »
Hello,

they look like TFK (Telefunken) Diodes.
Assuming the color code is grey+green it should be a BYW85.
Replacement according to data sheet would be a 1N5627.

Edit:
depending on wire diameter (0.85 mm) it could also be a BYX85.
The BYW85 should have 1.35 mm.

with best regards

Andreas
« Last Edit: August 10, 2022, 02:53:12 pm by Andreas »
 
The following users thanked this post: Vilnis_Radio

Offline Zenith

  • Frequent Contributor
  • **
  • Posts: 322
  • Country: england
Re: Vintage diode identification
« Reply #3 on: August 10, 2022, 02:35:05 pm »
What sort of welding machine is it - MIG, TIG or an arc inverter?

What brand  and model is it? There may well be a service manual on the WWW which would identify exactly what those diodes are and help you find NOS or current production equivalents.

The only diodes in my collection which look anything like that are Zeners, but yours can't be that.
 
The following users thanked this post: Vilnis_Radio

Offline Vilnis_Radio

  • Contributor
  • Posts: 6
  • Country: lv
Re: Vintage diode identification
« Reply #4 on: August 10, 2022, 03:07:12 pm »
Thanks guys,

The machine is Dalex-Werke make (German company that still doing good), VDE 0542, Type CGL 160. It is MIG welder. Well, I have searched for the diagram or some other information but without luck. It is vintage machine nowadays so no surprise that there is no info online. I even wrote an email to Dalex company but no reply. So far I narrowed the problem to minimum and I doubt that I could get some info on these diodes on the particular make. My guess that these are quite common diodes of that era.

Unfortunately I do not have square generator and I am not sure that I would be able to measure with certainty even if I had it. Looking at the machine I believe that diode must be slow :)

Yes, I also came across to  these Telefunken BYW85 but was not sure at all on that. Are avalanche diodes used in diode bridges? Initially I thought it should be very common diode. I found out that my local supplier could deliver me Vishay 1N5627-TAP. Does this "TAP" extension matters? Probably it is worth trying to use these 1N5627 and see what happens next.

Well, the wire diameter on original diodes is only 0.3 mm (not even 0.85). Actually 1.35mm is a really beefy one. Yeah, and the body of 1N5627 is 4mm while mine is about 2.5mm. Hmm, 1N5627 seems to be something other than original ones.

Nope, I rechecked with ordinary caliper and the wire diameter is 0.6 no 0.3 as I stated before. The body is somewhat 4.5mm and diameter 2.7mm. I used digital one before and realized that it can not be 0.3. :)
« Last Edit: August 10, 2022, 03:49:07 pm by Vilnis_Radio »
 

Online fzabkar

  • Super Contributor
  • ***
  • Posts: 1527
  • Country: au
Re: Vintage diode identification
« Reply #5 on: August 10, 2022, 04:50:27 pm »
Try researching the databooks of that era:

https://mirrorservice.org/sites/www.bitsavers.org/components/
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 14925
  • Country: us
  • DavidH
Re: Vintage diode identification
« Reply #6 on: August 11, 2022, 12:32:02 am »
Now I came to conclusion that the problem is absence of correct replacement diode and UF4007 is not the way to go. Circuit contains bunch of these diodes looking to be identical. So I put the rectifier instead of diode bridge to get spare diodes. When I changed UF4007 to original ones circuit starts to work as it should. The problem is that I need more of those diodes to completely change the bad looking ones.

UF4007's behave strange. In some point of circuit they seems preventing forward current. In some other point of circuit it leads to increased current and overheating cap and resistor. In the diagram D16 was one that overheated R1 and C2. D6 was blocking current to relay U5. When I replaced back to original ones both issues disappeared.

I would have expected UF4007s to work fine.  Could one of the replacements have been bad?

Yes, I also came across to  these Telefunken BYW85 but was not sure at all on that. Are avalanche diodes used in diode bridges? Initially I thought it should be very common diode. I found out that my local supplier could deliver me Vishay 1N5627-TAP. Does this "TAP" extension matters? Probably it is worth trying to use these 1N5627 and see what happens next.

Avalanche diodes are more reliable than standard diodes when there is the possibility of reverse breakdown.  Their junction is manufactured with greater uniformity so that breakdown occurs evenly without producing hot spots.  I think they would be used here because of the high probability of high voltage spikes.

I think the -TAP ending relates to how the diodes are packaged for delivery.  It is not relevant unless you are using them in a production environment.
 

Offline Vilnis_Radio

  • Contributor
  • Posts: 6
  • Country: lv
Re: Vintage diode identification
« Reply #7 on: August 11, 2022, 05:51:27 am »

UF4007's behave strange. In some point of circuit they seems preventing forward current. In some other point of circuit it leads to increased current and overheating cap and resistor. In the diagram D16 was one that overheated R1 and C2. D6 was blocking current to relay U5. When I replaced back to original ones both issues disappeared.

I would have expected UF4007s to work fine.  Could one of the replacements have been bad?



No, replacement diodes came new from supplier. Later on to prove my assumption I resoldered original, then put another UF4007 and then back original again.

Actually I am not sure about the voltage spikes on board. Board itself is fed by devoted transformer from mains with 48VAC. The only weird thing about transformer that it is supplied by two phases (not phase and null). All in all it that is ok however rarely used solution. Control board only moves several relays which in turn sets high magnetic switch, gas valve and DC motor for wire supply. Any welding currents are separated completely from control board.

Maybe some clue can give simple standalone circuit from main diagram. In the diagram D16, R1 (30 ohms), and C2 (220mf) together act as simple and rough DC from incoming AC converter. When welding machine is ON this part of diode, cap and resistor is ON too - waiting for the switch to close on my welding gun (when it does - three relays are turned on). So in this standby mode if there is original diode then ammeter shows around 10mA and both cap and resistor is cool enough. When I replace this diode with UF4007 ammeter shows around 200mA and the reading is growing up to 300mA, resistor and cap is overheating and I switch off the circuit.
« Last Edit: August 11, 2022, 06:10:26 am by Vilnis_Radio »
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 14925
  • Country: us
  • DavidH
Re: Vintage diode identification
« Reply #8 on: August 11, 2022, 08:13:30 am »
Actually I am not sure about the voltage spikes on board. Board itself is fed by devoted transformer from mains with 48VAC. The only weird thing about transformer that it is supplied by two phases (not phase and null). All in all it that is ok however rarely used solution. Control board only moves several relays which in turn sets high magnetic switch, gas valve and DC motor for wire supply. Any welding currents are separated completely from control board.

Anytime high currents are being switched, there is potential for conducted and induced high voltage spikes.

Quote
Maybe some clue can give simple standalone circuit from main diagram. In the diagram D16, R1 (30 ohms), and C2 (220mf) together act as simple and rough DC from incoming AC converter. When welding machine is ON this part of diode, cap and resistor is ON too - waiting for the switch to close on my welding gun (when it does - three relays are turned on). So in this standby mode if there is original diode then ammeter shows around 10mA and both cap and resistor is cool enough. When I replace this diode with UF4007 ammeter shows around 200mA and the reading is growing up to 300mA, resistor and cap is overheating and I switch off the circuit.

I do not understand how that could happen unless the replacement diode was bad.  There is very little practical difference between a UF4007 and the original diode.
 

Offline Vilnis_Radio

  • Contributor
  • Posts: 6
  • Country: lv
Re: Vintage diode identification
« Reply #9 on: August 11, 2022, 06:18:54 pm »

Quote
Maybe some clue can give simple standalone circuit from main diagram. In the diagram D16, R1 (30 ohms), and C2 (220mf) together act as simple and rough DC from incoming AC converter. When welding machine is ON this part of diode, cap and resistor is ON too - waiting for the switch to close on my welding gun (when it does - three relays are turned on). So in this standby mode if there is original diode then ammeter shows around 10mA and both cap and resistor is cool enough. When I replace this diode with UF4007 ammeter shows around 200mA and the reading is growing up to 300mA, resistor and cap is overheating and I switch off the circuit.

I do not understand how that could happen unless the replacement diode was bad.  There is very little practical difference between a UF4007 and the original diode.

Well it seems I have to play around a little bit with that. However I do not have "fresh" UF4007 to do it at the moment. All in all from other contributors I've got an idea that I might have to look towards avalanche diode in parallel. So far it is a bit confusing that I do not know for sure what particular avalanche diode I have to get as here suggested 1N5627 seems to be physically too big.
 

Offline wasedadoc

  • Regular Contributor
  • *
  • Posts: 187
  • Country: gb
Re: Vintage diode identification
« Reply #10 on: August 11, 2022, 06:44:25 pm »
What current rating is the fuse F1?
 

Offline Vilnis_Radio

  • Contributor
  • Posts: 6
  • Country: lv
Re: Vintage diode identification
« Reply #11 on: August 11, 2022, 06:47:00 pm »
Figured out that smallest axial avalanche diode body is SOD-57 with the wire diameter 0.82mm that is a bit thicker than my original with 0.6mm.
Here is the list with plenty of them (SOD-57). If I should get some for the tests what would be your suggestions for Vr/Io/Vf?
https://www.digikey.com/catalog/en/partgroup/avalanche-diode/11699
 

Offline Vilnis_Radio

  • Contributor
  • Posts: 6
  • Country: lv
Re: Vintage diode identification
« Reply #12 on: August 11, 2022, 07:00:50 pm »
What current rating is the fuse F1?

I would not rely on that fuse rating because it is obviously put in some time before I've got this machine and after it was sold. It has two post factory repairs (one old style reed relay changed to little bit newer type and this fuse).  It is not an initial factory circuit setup. Whoever put that fuse into the circuit meant to protect DC bridge and DC wire supply motor from variable transformer AC supply (current sensing transformer that adjusts motor speed with the welding current). The fuse has marking T4/250D that seems to be 4 amp fuse.

However there are original 4 fuses @ 2 amps on machine case. No doubt that they are meant to be 2amps because they all have clear factory marks on dashboard. One I see is for magnetic power switch (switches welding transformer to the 3phase mains) that is controlled by my control circuit. The others are hard to trace right now. Wire feed DC motor has following markings on its case - 39V, 52W and might be 1.9A (the amp reading is hard to see - have to look through mirror, but definitely below 2amps, 1.something). So that leads to 2amps all in all and 4amp fuse seems to be wrong choice here.
« Last Edit: August 11, 2022, 07:30:22 pm by Vilnis_Radio »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf