need help with 6DJ8 cascode build

[imaradiostar]

I've built an amp with a 6dj8 cascode out front, a simple tone stack and a fender style LTPI. It's built in a steel marshall valvestate chassis (the original amp blew up). Looking at the amp from the front the power transformer (an antek pt) is at the right, power tubes next to the left, vibroverb OT next, then preamp tubes at the far left. All tube sockets are bolted to L shaped metal brackets.

Here's the problem- dc leakage somewhere and ground hum and noise. I haven't tried to measure for leaky caps or bad gorunds with an o-scope but using voltmeters I can't seem to find a bad ground or leaky cap anywhere.

The pots are scratchy and just touching the ground lug at the preamp end changes the hum and scratching. The hum goes away completely when I touch the preamp end ground lug. Connecting an alligator clip to that same lug and grounding it anywhere (literally anywhere) on the amp quites things quite a bit. Measuring that same lug to ground without the alligator lead shows exactly 0 ohms. What the heck? All the grounds are solid. I've rebuilt the thing with star grounds, preamp grounds at the left and power all to one lug at the right, input ground at the left, pre in the middle and power at the right and every combo outside that I could think of. I soldered a 16 gauge wire to the back of the pots and grounded it just in case the pots weren't grounded correctly.

I recycled a tagboard from an old oscilloscope and I suspect DC leakage from the ancient board. I'm going to build a new turret board tomorrow to eliminate the recycled board as the problem if it is...but I doubt it.

I don't understand it. I've rebuilt the amp and rearranged the grounds many times. I'm wondering if it's because of the cascode or something. It's not like I've never built a dead-quite amp in a steel chassis before. It's extremely frustrating!

On the other hand the 6dj8 cascode sounds better than any previous 12ax7 cascodes I've built. I chalk it up to having tons of gain with a much lower plate impedance. Route 66 style tone controls, LTPI into cathode biased 6L6's- really makes for a neat sounding amp.

OK, done venting now.

jamie

[tubeswell]

Scratchy pot = DC on the pot or dirty pot. Could be bad coupling cap in front, or bad pot (splattered resin between the pot terminals, or strained pot lug somehow)?? Test the cap for DC leakage or try cleaning/replacing the pot.

Hum ?? grounding hum or heater hum or filtering hum?

[imaradiostar]

Well- it's not the recycled o-scope board. I built a nice little turret board to replace the old board and it made no difference. It's a lot prettier though!

I haven't tried replacing the pots yet but I will tomorrow. I hate it when I use all brand new parts then rip them out a day or two later and put other brand-new parts in which do exactly the same thing.

Maybe I'll take some digital pics and post them- this thing is not a normal build by a long shot. I still need to get in there with a higher impedance measuring tool (o-scope) and see if I can figure out why it's acting up. Hopefully I'll be able to see any ground differential with the scope.

jamie

[imaradiostar]

OK, see if this makes any sense at all- after probing with the o-scope and only finding a little rectifier hash at the grounds (no unusual DC or AC) I was poking around and noticed the noise went away as the plate voltage dropped when I flipped it off. I increased the dropping resistor for that stage from 1k (around 380 volts) to 15k (now around 340) lowering the voltage to the cascode and now it's quiet.

what the heck?

any thoughts?

jamie

[loudthud]

Could the circuit have been oscillating above the bandwidth of your scope? That's one thing I think of when things don't make sense.

[imaradiostar]

would oscillation show up as voltage of some sort or would I just see nothing?

I have a 20mhz scope I think...have to double-check.

jamie

[imaradiostar]

Nevermind- I have a tektronix 5110- 2mhz.

jamie

[imaradiostar]

I should be more clear- the pots are still bad and scratchy but the input hum is gone. There is no DC on the pots so I must have gotten them dirty or something. Very strange. I'll replace them and see if it makes a difference.

jamie

[loudthud]

Tek 5110? I used to test high voltage/deflection boards for that scope in the 70's.

[imaradiostar]

Originally Posted by loudthud

Tek 5110? I used to test high voltage/deflection boards for that scope in the 70's.

so you may have built my scope?

jamie

[loudthud]

I just tested the HV board, adjusted the HV to -3450V, checked that the deflection amps were working, and I think a couple of other pots on the board. IIRC I put an X on the little power transistor bracked where you can't see it when the scope is assembled.

[Steve Conner]

Could it have been heater-to-cathode leakage on the top section of your 6DJ8? Try elevating the heaters.

Also, make sure you're using the sections the right way round: the two sections of a 6DJ8 are optimized for their respective places in the cascode. The top one may well have better heater-to-cathode insulation, and the bottom one lower noise.

If you have an AC coupled cascode where both cathodes are at low potential, ignore me.

Loudthud, I've got a Tek R7603, did you build that one too?

[imaradiostar]

Originally Posted by Steve Conner

Also, make sure you're using the sections the right way round: the two sections of a 6DJ8 are optimized for their respective places in the cascode. The top one may well have better heater-to-cathode insulation, and the bottom one lower noise.

I'm pretty certain I used them in the appropriate grounded cathode and grounded grid positions but I guess I could have gotten it backwards. I need to double-check!

The output section is cathode biased- I should move the heater CT resistors to one of the output tube cathodes.

jamie

[loudthud]

My first job at Tek was in this group that did overflow assembly work for other product groups. About 95% was just PCB assembly. The only things I remember from the 7000 series were readout boards, this little vertical channel switch board, and occasionally deflection boards. It was a part time gig but you could take classes Tek offered in electronics.

The R7603 is a great scope (I have a several 7600 series mainframes) if you have the right plug-ins. 7A22... 1 micro volt per div @ 1MHz and 7A13... 1 milli volt per div differential with DC comparitor. 7A18 and 7A26 nice for general purpose dual trace work.