HP 6114A repair help.

[TK0130]

So my HP 6114A recently gave up. Had it just running on the desk powering a switching regulator I was experimenting with and it fizzled. (Let out a ton of magic smoke while it was at it...) I have had this brick for about 6 years. I saved up mowing lawns for quite some time before purchasing it at a hamfest. At the moment I feel like I am back in the stone ages with a computer ATX supply and some home made linear supplies.


So onto the repair. When I turn it on I can hear the transformer humming. the meter shows above max current (2.4A), and the overcurrent light is on.

The magic smoke all came from a 1 resistor on the board. It is connected between ground and the emitter of one of the four transistors(? more on this in a moment) on the rear heatsink. Pulling the TO-3 package off the heatsink it simply states "V0 7423452 RCA" I instantly get a feeling of dread as this looks like an in-house part number. Google was no help at all. Testing it, the "can" and the "E" pin are shorted and the other pin has a ~0.4v drop to the case and the "E" pin. I am assuming that "E" on the pcb stands for "emitter" as the same pin is labeled likewise on all three other TO-3 cans.

It is here that I hit a standstill as I cannot find a replacement for the part. Are there any suggestions as to a possible replacement? Could it just be a rebranded 2N3055 (or similar) as RCA is known for doing a custom number for a single product.

Schematic of similar model: (Snitched from here: https://www.eevblog.com/forum/reviews/would-you-pay-$100-for-a-hp6115a-psu/msg284891/#msg284891 )


Wondered if the SCR is kicking in and shorting the output. Pulled the drive resistor for it (R27, 47 ohm) and nothing changed. Put it back in.
Replaced the totally smoked 1ohm resistor (R10, 1ohm 5%) with 10 10 ohm resistors in parallel (all I had around.) and now when I turn it on it doesn't ping to full power on the meter. I can adjust the current limiting knob and make the meter movement change from between 0 and 0.2A.

Thanks,
Nathaniel.

Oh, and is $140 a good price for a used 0-300v 0-3A DC power supply? The ripple is quite bad rated (700mv at max voltage) but it has constant current and voltage. (Yes you read that right. 300V DC.)

[scl]

Let's start with making sure we know which transistor is shorted.
The full manual is at keysight: https://www.keysight.com/main/redirector.jspx?action=ref&cname=EDITORIAL&ckey=734221
Look at Figure 7-1 to see which heatsink transistor is which. If it is Q2/Q3, the manual says those can be replaced with 2N3442, and Digikey has those in stock for a bit over six bucks. I don't see a 1 Ohm to ground from any of the four transistors on the heatsink, so I'm confused there, but the schematic does show 1 Ohm resistors to Q2/Q3 emitters so thus my guess it is one of those.

[TK0130]

After some searching I found a real service manual. It lists the power transistor as a 2n6254 for Q1-3 and a 2n5954 for Q4. Looks like I will be ordering some.

[TK0130]

Let's start with making sure we know which transistor is shorted.
The full manual is at keysight: https://www.keysight.com/main/redirector.jspx?action=ref&cname=EDITORIAL&ckey=734221
Look at Figure 7-1 to see which heatsink transistor is which. If it is Q2/Q3, the manual says those can be replaced with 2N3442, and Digikey has those in stock for a bit over six bucks. I don't see a 1 Ohm to ground from any of the four transistors on the heatsink, so I'm confused there, but the schematic does show 1 Ohm resistors to Q2/Q3 emitters so thus my guess it is one of those.

But the line items in the manual says that substitution is for the 6105A and 6115A. However, it does look suitable. I was looking at a 2n3773G from ON semi on mouser.

Before I had a good schematic and manual I mesured from the resistor that fried (R10 I think it was) to ground and it registers as a short (0.6 ohms). But now I realize that is because the outputs are shorted together and I am reading the current sampling resistor (R15, 0.5 ohm).

Something else is really wrong so I might as well dig some more before ordering those transistors to save on shipping.


Nathaniel.

[coromonadalix]

oops you posted before me  loll

Yeah  it depends on the model you have

You can disable the ovp circuitry by removing cr20  for tests
 
Check al the other voltages like the A2 control board  +16vdc  and -6.3vdc,  the biais supply section too  +7.5 vdc and 3vdc, and the constant current source section. You have +28 and -28vdc, and -4.3vdc.

The circuit ground is the transformer pin 16, the other secondaries are floating voltages.

Check Q1  outside the circuit,  all the psu power pass thru this one ??  strange design,  somekind of extreme current failsafe if the current limitter doesnt work or do well,  called gross current limit.

R10 to ground mean the ovp is shorted out, or there is an isulation problem of some sort ?? or CR15 is bad / leaking / shorted out  , its an reverse protection diode.

Has anyone played in it before you did bought it at the Hamfest ?

You could add reverse voltage protection on your Q2 Q3  power transistors by adding a fast recovery power diode,  cathode on colletor, anode on emitter.



Oh and finally  is your rear screws "jumpers" terminals are at the right place,   but if the psu worked fine before, they should be at the right place  ex :  the sense lines wires are connected at the two main outputs plug / terminals

[scl]

Yes, dangit, I slipped a line on the BOM, you're right, need a 2N6254. Looks like Jameco has those if you want to stick with the same part.

I see coromonadalix has plenty of leads for finding the short, so you're in goods hands.

[TK0130]

oops you posted before me  loll

Yeah  it depends on the model you have

You can disable the ovp circuitry by removing cr20  for tests
 
Check al the other voltages like the A2 control board  +16vdc  and -6.3vdc,  the biais supply section too  +7.5 vdc and 3vdc, and the constant current source section. You have +28 and -28vdc, and -4.3vdc.

The circuit ground is the transformer pin 16, the other secondaries are floating voltages.

Check q1  outside the circuit,  all the psu power pass thru this one ??  strange design,  somekind of current failsafe if the current limitter doesnt work or do well,  called gross current limit.

R10 to ground mean the ovp is shorted out, or there is an isulation problem of some sort ??

Has anyone played in it before you did ?

Not too concerned by OVP as it doesn't seem to be triggering. May disable it anyway as that is one less thing to worry about.
First commandment of electrical repair: "Thou shalt check all voltage rails!" all of them are dead on.
Have been measuring all voltages in relation to the negative output since it is connected directly to pin 16.

I will check Q1 but I think it is worth tracking down why R10 seems to be shorted to ground.



*5 minutes later*

On a whim I pulled CR15 and it is fine. So I pulled the wire from the main PCB to the + output terminal, still shorted. So that is odd... So the electrolytic failed short. Still not the only problem but one less to deal with. Pulled C13 (15 (uF?) 50V) and sure enough it is shorted. Let's see what it does now.


Nathaniel.

[TK0130]

Voltage regulation is working perfectly now. Still current limits at ~0.23A. Now to take a look at Q1.

[TK0130]

Q1 checks out fine out of circuit.
Q2 ~0.5V drop from B to E and ~1.5v drop from E to B out of circuit.
Q3 see earlier post.

Looks like Q2 and Q3 are gone. Tested everyhing else and it seems to be fine. But why is the power supply voltage regulating if the transistors are shot?


Missed a question a few posts ago: No one has worked on it before. I can tell I am the first person to tinker around inside this thing. All the solder joints look original. It even had the original fuse in it i believe.

[coromonadalix]

What is the voltage at the psu output    if q2 and q3 are short you should have  maybe around 40 volts at the output   and Q4 driver maybe busted too.


You can make a test with new q2 and q3 (maybe an 2n3005 if you dont go over 60volts) :  unsolder cr3  it will isolate the current control u3 ic,  r15 needs to be there and not open.

If your voltage control doesn't work,  you still have some troubles ahead

I think you had an huge reverse voltage in your psu, it busted the main transistors ...

Have you checked the vr3 vr4 zenners ?

Have you checked the Q1 and cr13, cr14, cr22  ?

[TK0130]

What is the voltage at the psu output    if q2 and q3 are short you should have  maybe around 40 volts at the output   and Q4 driver maybe busted too.


you can make a test with new q2 and q3 (maybe an 2n3005 if you dont go over 60volts) :  unsolder cr3  it will isolate the current control u3 ic,  r15 needs to be there and not open.

If your voltage control doesn't work,  you still have some troubles ahead

I think you had an huge reverse voltage in your psu, it busted the main transistors ...

Have you checked the vr3 vr4 zenners ?

The weird thing is that voltage regulation works. It outputs exactly what it is set to. I don't understand it. I don't want to toss in some 2n3055s (Mostly due to the fact I only have one on hand. I have a ton of mosfets ) and the fact that the bus voltage is near it's limit. I will get some power NPNs. Do you think 2N3773 will work?


Due to the reverse voltage diode being intact I think it must have been overvoltage. (I was driving a highly inductive circuit)


Edit:
Removed CR3, still current limits at about 0.23A. Current limit LED lights up.

[coromonadalix]

An 2n3773 would do perfectly to replace Q2 Q3,  if you cant output more than 0.23 amps  its definitly wrong

Remove Q1, bypass the emitter collector, and bypass R6,   this will remove the extreme current protection

With CR13 removed you shouldn't  have any current limit left/ active.

Once again,  is the voltage regulation work with q2 or q3 substitution ?, what are the value you get, can you drop to zero and get to the max voltage ?,  you should only put  q2 or q3  not the two  to do tests.

Ex : if the voltage varie from 0 to 40v the voltage regulation is working perfectly, unless you have some problems in the decade resistos section to select your output voltages ?

Next :  find an power resistor of 8 ohm 10w, adjust to 8 v at the output and check if you have around 1 amp at the output,  ex 16 ohms load  20 watts for 16 v at the output is still 1 amp.

Dont do it for too long,  we need to be sure the regulator section is working.

Dont short it directly with an ammeter / multimeter,  an ammeter has an lower impedance than 1 ohms.

The minimum  decade resistor values  between x and y points  is 20 ohms up to around 200k ohms  (if you add  all the ohms values)

And your screws post terminals must be like the photo  "the green boxed terminals section". 

I ve bought an kepco ate 75-5  with the missing terminal block for almost nothing,  but high shipping  fees for an 57 pound  psu  loll , The guy did not knew it was needed to work loll

[TK0130]

An 2n3773 would do perfectly to replace Q2 Q3,  if you cant output more than 0.23 amps  its definitly wrong

Remove Q1, bypass the emitter collector, and bypass R6,   this will remove the extreme current protection

With CR13 removed you shouldn't  have any current limit left/ active.

Once again,  is the voltage regulation work with q2 or q3 substitution ?, what are the value you get, can you drop to zero and get to the max voltage ?,  you should only put  q2 or q3  not the two  to do tests.

Ex : if the voltage varie from 0 to 40v the voltage regulation is working perfectly, unless you have some problems in the decade resistos section to select your output voltages ?

Next :  find an power resistor of 8 ohm 10w, adjust to 8 v at the output and check if you have around 1 amp at the output,  ex 16 ohms load  20 watts for 16 v at the output is still 1 amp.

Dont do it for too long,  we need to be sure the regulator section is working.

Dont short it directly with an ammeter / multimeter,  an ammeter has an lower impedance than 1 ohms.

The minimum  decade resistor values  between x and y points  is 20 ohms up to around 200k ohms  (if you add  all the ohms values)

And your screws post terminals must be like the photo  "the green boxed terminals section". 

I ve bought an kepco ate 75-5  with the missing terminal block for almost nothing,  but high shipping  fees for an 57 pound  psu  loll , The guy did not knew it was needed to work loll

Removed Q1 and jumpered R6.

Output is now ~21.5v and doesn't react at all to voltage adjust. I assume this means that q2 and q3 are shot. Q1 seems to have been the only thing doing any work.

Pulled Q2's emitter, same thing.
Pulled Q3's emitter and the output goes to ~1.2v (from voltage divider at current adjust i guess?)

I guess as I pulled Q2 Q3 and Q4 and they all tested odd I should replace all of them. Anything else I should check before I go and put new ones in?
Och that PNP is expensive I don't see an easy drop in replacement on mouser/digikey. any input? closest I can find is the 2N3792.

Oh and I have a line on a 0-300v 0-3A raytheon/sorenson power supply for $100 or a 0-80v 0-16A one for $100. I would love the high current for some motor drive stuff but I would also like the 300V to mess with some vacuum tubes. are they worth it and which one should I get? note that the ripple under load of the 300v one gets quite bad at full power. (0.7v ripple) while the 80v one is only 20mv.

[coromonadalix]

It depends of your needs,  check the service manuals / schematics for them  ??

A 80v-16 amps would be ideal for me  loll   Is it an Kikusui  ??    Kikusui can send some service manuals  under NDA ...  i have one for all the PAD series.

I dont think an 300v for tubes application would help,  my friend who repairs and restore tube radios says:   you should need at least 450 to 550 vdc

Lambda services manuals are rare, i dont know or posess one
Sorensen DCR series have services manuals available, i have the DCR series, they exist on the web.
Kepco ATE series, i have the service manual for them,  hard to find,  but they exist on the web.
Xantrex,  some series or almost all the rackmount series service manuals aren't available.


Can send you copies if you want ?

80vdc at 60 amps on Ebay for 100$ ??  I think its an Lambda in disguise ?
https://www.ebay.com/itm/Electronic-Measurements-EMS80-60-DC-Power-Supply-0-80V-0-60A/352545168930?hash=item5215545e22:g:~yMAAOSwj5BcFEWJ:rk:16:pf:0

[coromonadalix]

For the 6114a

If you dont get over 22vdc  you may have an other problem in the regulation op amplifier circuit ??

Your psu seems to have busted pretty severly

Maybe you should move on to another one psu ??  just saying ?

[TK0130]

It depends of your needs,  check the service manuals / schematics for them  ??

A 80v-16 amps would be ideal for me  loll   Is it an Kikusui  ??    Kikusui can send some service manuals  under NDA ...  i have one for all the PAD series.

I dont think an 300v for tubes application would help,  my friend who repairs and restore tube radios says:   you should need at least 450 to 550 vdc

Lambda services manuals are rare, i dont know or posess one
Sorensen DCR series have services manuals available, i have the DCR series, they exist on the web.
Kepco ATE series, i have the service manual for them,  hard to find,  but they exist on the web.
Xantrex,  some series or almost all the rackmount series service manuals aren't available.


Can send you copies if you want ?

80vdc at 60 amps on Ebay for 100$ ??  I think its an Lambda in disguise ?
https://www.ebay.com/itm/Electronic-Measurements-EMS80-60-DC-Power-Supply-0-80V-0-60A/352545168930?hash=item5215545e22:g:~yMAAOSwj5BcFEWJ:rk:16:pf:0

Sorry misquoted it. 80v 12A sorenson DCR80-12B. I would defiantly want to build a secondary regulator. Interestingly it has a datasheet and repair sheet but the main device is a thyristor quoted only as "S12" so I am wary of buying it as I cannot get replacement parts for the main switching device.

Shipping is what kills me. I don't want to spend over $150.


As to my supply: I really want to fix it as one in this condition can go for $300 or so on e-bay at times. Mine is in near mint condition (Other than someone rivited a label to the back...) and was accurate to 0.01% across it's entire range. Great for high precision stuff.

[coromonadalix]

The sorensen 80-12b  has an 2n5445 40 amps triac  for CR7, this is the main device, easy to substitute.   Yeah  i know  they are heavy as hell, big shipping fees

And what do you know  loll there's a few lm723 in it  lolll


Your 6114  is still repairable,  you need to work the section one after the other,  the voltage regulator section seems not working well,  you have to redo some checks,   change the op amp  U2  and check his -14.7 vols and the +16 volts and CR2

Check the X and Y  voltage decade switch section, as i said it will be minimum 20 ohms and up to 200k ohms when playing with the encoders.

[TK0130]

The sorensen 80-12b  has an 2n5445 40 amps triac  for CR7, this is the main device, easy to substitute.   Yeah  i know  they are heavy as hell, big shipping fees

And what do you know  loll there's a few lm723 in it  lolll


Your 6114  is still repairable,  you need to work the section one after the other,  the voltage regulator section seems not working well,  you have to redo some checks,   change the op amp  U2  and check his -14.7 vols and the +16 volts and CR2

Check the X and Y  voltage decade switch section, as i said it will be minimum 20 ohms and up to 200k ohms when playing with the encoders.

Well just shoot me now... The 16v rail is at the full 32v... This means U1, U2, U3 are shot along with Q2, Q3 and Q4. At this point I might as well pull every semiconductor and test them all then replace them all... The odd thing is that I checked all the voltage rails and remember them being dead on. Thus it was either delayed damage due to shorted transistors or I caused it in some way.

Looking at the ratings of many diodes I doubt they are damaged. We shall see with the zeners...

At this point I am going to get that sorensen so I have some supply in the meantime.


Pricing it out, only the known dead parts come out to $60 before shipping. That is getting dangerously close to the list price of some units in poor aesthetic condition. I am starting to think this is a NER. (Non economical repair) As much as I hate to say it I may have to break down and buy a new supply.


Any chance you could divulge where you found the info on that triac? I looked around for quite some time and never figured out a replacement.

[james_s]

Last time I had a power supply that the current limit was way too low the problem was the current sense resistor, I assume you checked that?

Also if it has TO3 transistors in sockets, check the sockets, I've had those get flaky a few times.

[TK0130]

Last time I had a power supply that the current limit was way too low the problem was the current sense resistor, I assume you checked that?

Also if it has TO3 transistors in sockets, check the sockets, I've had those get flaky a few times.

If only it was that easy. Nothing is socketed and the sense resistor is fine. The 16v rail is at 32+ volts...

[coromonadalix]

Not so sure   
Be careful of the interpretation, it may indeed be +32v

BUT you have floating supply (ies) in there

For the A2 control board : for the +16v, -6.2v and -14.7v  the common will be at the "+S"  output terminal,  with the "+"  main output terminal  (screw post #2  #12 and #10).

Not necessarly the transformer pin 16 ??

[TK0130]

Not so sure   
Be careful of the interpretation, it may indeed be +32v

BUT you have floating supply (ies) in there

For the A2 control board : for the +16v, -6.2v and -14.7v  the common will be at the "+S"  output terminal,  with the "+"  main output terminal  (screw post #2  #12 and #10).

Not necessarly the transformer pin 16 ??

  This is why you don't stay up until 2am working on stuff like this. must not have looked at the schematic at the time.

28v and -28v are about 29.5 but as they are direct unregulated rails I am not concerned. Everything else is dead on (So says my fluke 117).

Resistor array (X and Y inputs) are fine.

I only have a pair of NTE941 opamps in the TO-99 package. do you think it would work? the original appears to be similar to an LM741 (almost identical ratings) Would it be worth making a hack board for a dip package?


I will pull CR3 and leave it out for now. Pull CR2 and test it CR2 and CR3 test good.
Replaced U2. No change. Old opamp appears to be good in some test circuits. Output (pin 6) of U2 is -12v in relation to +S and 8v in relation to ground (- output). Other side of CR2 (junction of CR2 and CR3) is at about that same voltage.

[james_s]

Yeah I have a tendency to do that as well, but it's usually a good idea to just call it a night before it gets super late. An added bonus is that if I go to bed thinking about it, I often wake up the next day with new ideas.

[coromonadalix]

For the op amps  you could put ic sockets and put your to-99 in them     had to do it  for replacing lm723 with an to-99 variant  in a socket  loll

[Tomorokoshi]

I just replaced Q2 and Q3 in mine with ON Semiconductor 2N3773G:
https://www.digikey.com/products/en?keywords=2N3773GOS-ND

Now it seems to work just fine. Q2 got burned out when I used it to supply a different device being tested at 2kV EFT at the EMC lab. Don't ask.

The original Q3 seems to measure fine, but it's best to replace them as a set.