Re: Output transformer for K5881 - disaster!

[Lyndon Johnson]

I'd appreciate some advice here!
I decided just to replace the faulty one on account of the fact that some chassis drilling was needed. This now resided in the previously good RH channel after my earlier diagnosis work.
After checking and double checking my wiring I put the valves in and switched on to the sound and smell of the 9 Ohms worth of resistors vapourising across the speaker terminals of the RH channel. Thinking they were just too low wattage I disconnected them (though noted the LH ones were cool).
Powered up again and this time noticed immediate smoke from under the base of one of the RH channel 5881's. ( actually Svetlana 6L6 GC).
At this point I decided to disconnect the V3 and V4 anodes from the OPTx and measure some voltages:
Left. Should be Right
OPTx CT 480v. 448v 479v
OPTx A1. 469v. 440v 478v ( not connected to V3)
OPTx A2. 468v. 440v. 478v. Ditto
V1A-R5. 107v. 97v. 79v
V2/1. 337v. 333v. 241v
V2/6 336v. 330v. 246v
V4/4. 441v. 420v. 320v

At this point I noticed that R17 ( HT to V3/4 screen grid) was overheating and that V3 was now missing it's silver at the top of the glass envelope.

Clearly V3 has failed and maybe this took out R17 but was this caused by my fitting of the new OPTx ?

Any advice on what my next step should be?

[John Caswell]

Hi Lyndon,
This is beginning to sound like parasitic oscillation in the output stage, or possibly feedback with the wrong polarity.
It is possible that you have wired it incorrectly, and it is also possible that the replacement transformer has been assembled incorrectly by the manufacturer.
Removing the anode feeds from the OPTX and then powering up will pretty certainly destroy the output valve(s) plus the associated screen grid resistor. Why did you do this?
Unless you are absolutely sure you know what you are doing I suggest that you leave well alone and find someone who is more experienced in this kind of thing. I suggest initially that you appeal to WADers in your area via this forum to see if they can help.

John

[Lyndon Johnson]

John
Thanks for responding. When you say wired incorrectly , do you simply mean The V1 and V2 connections the wrong way round?

This is my backup amp. I can afford to take my time and I am happy to learn from my mistakes, provided I keep safe.

I will Replace R17 ( I have a spare as the LH one failed about a year ago prior to a valve failure). Then fit my trusty old Sovtek 5881s . I will swap V2 and V1 round on the OPT. The manufacturer did comment that I should do this if I got oscillation.

Lyndon

[John Caswell]

Hi Lyndon,
First thing is replace the screen grid resistor R17, then reconnect the anode feeds to the transformer. Fit a dummy load to the output sockets (8R ish ) plus any old rubbish loudspeaker in parallel with the dummy load. Connect up, power on and wait for it to warm up. At the slightest hint of oscillation/motor boating from the L/S power down and then swap over either the OPTX secondary leads to the LS sockets OR the OPTX primary anode leads, NOT both together.
Try again, if you still have oscillation/motor boating then you have further problems that require more brain power/diagnosis.

John

[Lyndon Johnson]

Thanks John,
I'll do just that and report back
Lyndon

[Lyndon Johnson]

Sorted thanks! I have a happy, bouncy amp again.

It all went as you suggested - oscillation then none when I reversed the anode leads. I know had wired them correctly initially because the swap required me to lengthen one of the anode leads to reach the OPT terminal.

So I have learned:
1) wire an old speaker in parallel
2) never run the amp with the power valve anode leads disconnected!

Thanks again John you're a star.

Lyndon

[John Caswell]

Hi Lyndon,
Excellent, well done!
Unless you have a 'scope or a **** LS on the output it is almost impossible to tell whether you have the feedback correctly phased, it can be apparent but not always. I would suggest that you also tidy up the wiring if necessary, as the 5881s were a bit prone to HF oscillation due to long anode leads. Try to keep the anode and grid leads apart.
If you disconnect the anode leads then the valve thinks the screen grids are anodes and draws a lot of current through them and goodbye output valve(s)

John

[Lyndon Johnson]

Thanks John.
Things are making sense now. I realise that the C17 burnout and valve failure that I suffered last year were most likely a consequence of the LH OPT (that I have now finally replaced) failing to deliver HT to the anode of one of the power valves.
Lyndon