[ruZZ.il]

Alright, maybe someone can help me out..
the entire left channel seems to be behaving, the tube voltage sets, and the bias is within range. The right channel, not quiet. With a tube inserted, the tube bias range is from 16~19.8v (Main voltage set to 22v), while the DB bias starts at around 180mV, and climbs to about 1.3V, before I promptly disconnect it all. I have a feeling the output transistors may be damaged since the trimpots were 'maxed' in the wrong way for a bit when I first powered up.. I found a local place with the least recommended BD139/140s that I'll pick up tomorrow and see how that works for now... meanwhile, maybe someone can spot something else?
Thanks.
-Russ.

Top-Large, Bottom-Large.

Thanks guys..

(the 3 wire braid is temporary, and the fuse blew so I bypassed

)

 

[tomb]

I don't see anything from the pics, ruZZ.il. It's immaculate work, by the way - outstanding.

It's possible the JFET gave out before any of the output transistors. 1/2 of the total current passes through that one tiny transistor. It's the weak spot in the DB, not the output transistors.

What are you biasing the left side at? 50ma is a proven current. Colin has stated 100ma shouldn't be a problem, but I'm not sure he's considered the consequences to that little JFET. (That's the PN4392 transistor - one per channel.)

EDIT: This guess is questionable. I'm not so sure of my facts and the voltage that JFET sees. It was an issue in the 1st prototype. So, we turned the current mirror into a current multiplier. That lowers the current seen by the JFET.

 

[tomb]

Also, even if 100ma is attainable, it's not with 1" high heat sinks. 50ma (110mV) bias is just about maximum, IMHO. The output transistors are hot enough at that size heat sink and current.

What you describe at 180mV is thermal runaway. The BJT's are positive temperature coefficient devices, so as temperature climbs, so will their current, ala voltage. So, they are burning up toward the 1.3V.

My first guess may be all wet.

 

[ruZZ.il]

I checked QB1L vs QB1R (with the diode check on my DMM, for forward bias voltage), and they measure up pretty much the same, but then so did qb8/9. I'll see if I can pick up one/two of those anyway. I haven't set my left DB bias yet, just to minimum till the right is sorted out too, but it starts at 22mv and biases up seemingly alright, so I'll leave it at that for now. I'll go over all my resistors and stuff again too.. anyway, just for fun heres a pic from my Adobe x-ray cam

Thanks.. and if you think of anything else I should ckeck let me know. (will be back home later, and am bound to be the night owl that I am...)

edit: also, any idea why one side tube bias wouldn't set right?.. would something in the DB section effect this?

 

[tomb]

Quote:

Originally Posted by ruZZ.il /img/forum/go_quote.gif edit: also, any idea why one side tube bias wouldn't set right?.. would something in the DB section effect this?

There's a few volts that drop on the way to the CCS and tube. The voltage difference you cite of 22 to 19.8 just sounds like it's not biased, period (can't speak for your range down to 16 - that sounds like it's "caught" and is starting to bias) ... and no, nothing in the DB section would affect this.

 

[ruZZ.il]

hm. my next option is to start switching components between L and R and see if I can follow the problem.. fun fun

any risk at powering up with a missing QB1/8/9 ?

 

[[AK]Zip]

What do you guys think about VitamineQ caps for CA3 and CA6 as well?

 

[tomb]

Quote:

Originally Posted by ruZZ.il /img/forum/go_quote.gif hm. my next option is to start switching components between L and R and see if I can follow the problem.. fun fun any risk at powering up with a missing QB1/8/9 ?

ruZZ.il, let's try a more cautious approach if you're up to it. To be honest, I'm going to be worthless helping anyone troubleshoot separate transistors. You are almost better off to just replace the whole bank of 2n5087's and 2n5088's (it's what I'd do and have done), but before you do that:

- Shoot for a bias of about 20-30ma to start. Despite what Colin suggested, no one that I'm aware of has tested the BJT's at higher than 50ma bias - and they're positively great at that point, anyway.

- You might go over all your resistors - make sure they didn't get switched.

- Turn the Left channel down to about 20-30ma. Then turn it off and measure the pins of the Left DB trimmer (ohms). Then try to adjust the Right DB trimmer for the same values. Turn the power back on and try it that way.

 

[tomb]

Quote:

Originally Posted by [AK]Zip /img/forum/go_quote.gif What do you guys think about VitamineQ caps for CA3 and CA6 as well?

Probably a waste - they're not in the signal path. Ferrari uses them or something similar, I think - maybe Russian teflons - but his extreme builds are not something to recommend for everyone. He'd probably say that, too.

 

[ruZZ.il]

tom, I did that already

(the trimmer thing) I've also checked my resistor values, and compared the resistance checks of each side. I've also switched QB1 between sides, and the problem stayed on the right. I'm in the process of switching QB8/9, will see if it follows.. otherwise I suppose I'll replace the rest of the transistors (good thing I got hundreds of those..)

 

[[AK]Zip]

Here is where I am so far.

I am just waiting on some better parts for QB8 and QB9 and some tube sockets. Volume controller will be a stepped attenuator. I really am thinking about changing out the Wima caps for the VitamineQ's for fun.

Also, thinking about installing the R1 instead of the jumper..

-Alex-

 

[Ferrari]

Yes indeed! For these positions you can best stick on the recommended MKP's as specified in the BOM (Wima, Ero MKP).
If you consider VitamineQ caps... you can best put them on C8-9 L/R, where the audio signal has the most benefit from them.

I put Russian PIO on C3 and C6 just because I have got the whole box of these caps for my birth day.
(A good friend knew that I'm a DIY junky

). For C8 I use Russian teflon and for C9 polystyrene (the yellow caps I used in my first Millett).

Ah... you posted when I'm writing this. I see that you already put VitamineQ on C8-9 L/R, very wise !

 

[ruZZ.il]

My problem followed QB8 from the right to the left, including the tube bias voltage. bugger. thanks again for the suggestions.. and I guess now we know a problematic QB8 can effect tube bias too now, anyone got a few spare output transistors to send me?

I'll pick up a few of the fairchild ones in the meantime.. but ordering from BDent is a problem for me ;/ on that note, if anyone is about to make an order and wouldnt mind adding a few in for me, and shipping them snail mail, I'll paypal you for it.

thanks..

 

[tomb]

ruZZ.il,
YGPM.

 

[babbkutz@comcast]

I'm using dayton RCA jacks from Parts Xpress....they have dual teflon washers (front and back) with a ground ring. I'm thinking washer/panel/washer/ground ring/nut......is this correct? or should the ring touch the case panel?