built-in dumbleator pwr question

[Tonegeek]

I am rethinking how I did the power to my onboard dlator.
does this look like a good starting point? B6 would go to the 100k resistor on the recovery plate and b7 directly to the cathode follower plate. The dlator string is parallel to the screen supply. Using Marshall 50w PT (I forget the voltage specs) Is the layout correct and the R values close for 220v and 260v respectively at the plates? I mainly wasnt sure about the 33k resistor position but thinking it is dropped off the 47k and not off the main string.

Thanks

[dogears]

Right, the string is series. I would absolutely add another node though. Before B+6 and B+7, I would tap into the screen supply and use a 33uf cap to then feed your string. Lower the B+6 47K to 24K. Add another 24K from the supply to decouple the 33uf that then feeds all of it. So 3 caps for the loop.....

[Bob-I]

The only problem I see is the R values. The 47K is too large. I used 22K and 2.2k on mine and it worked out great. I took the branch from the same location and used 22uF's just like yours.

[Fischerman]

If you want to do some 'ballpark' calculations...each triode should pull in the 1mA neighborhood.

[dogears]

Hey Bob,

I used very similar values to Winston's on my amps. Voltages are perfect. I'd have far too much plate voltage with only 22K there. Maybe you are powering cathode then plate?? Real Dumbleators power plate then cathode.

The only problem I see is the R values. The 47K is too large. I used 22K and 2.2k on mine and it worked out great. I took the branch from the same location and used 22uF's just like yours.

[Funkalicousgroove]

you want to see 257 vDC on the plate of the cathode follower side, and you want to see 368 vDC before the 100K plate load resistor, and 229 vDC on the plate itself. the real article uses a 100K dropping resistor between 2 33uf filters.

[Bob-I]

Hey Bob,

I used very similar values to Winston's on my amps. Voltages are perfect. I'd have far too much plate voltage with only 22K there. Maybe you are powering cathode then plate?? Real Dumbleators power plate then cathode.

Oops, this is the 6V6 amp with a small B+ voltage.

[Tonegeek]

Right, the string is series. I would absolutely add another node though. Before B+6 and B+7, I would tap into the screen supply and use a 33uf cap to then feed your string. Lower the B+6 47K to 24K. Add another 24K from the supply to decouple the 33uf that then feeds all of it. So 3 caps for the loop.....

Scott, Did I translate what you said correctly? (see image)

[Fischerman]

I'm pretty sure that's what dogears meant tonegeek...he suggested this to me as well...unfortunately I don't have any more room in the doghouse (I'm using a Fender chassis). Just play with the resistor values to get your voltages right.

Not to hijack too much but...why are there two nodes for the loop? I know the Dumbleators were like this but I've put them in a few of my amps just using one node and they work fine, seem quiet, and don't oscillate. Using one node yields higher plate voltages on the CF but that hasn't been a problem.

[Bob-I]

I'm pretty sure that's what dogears meant tonegeek...he suggested this to me as well...unfortunately I don't have any more room in the doghouse (I'm using a Fender chassis). Just play with the resistor values to get your voltages right.

Not to hijack too much but...why are there two nodes for the loop? I know the Dumbleators were like this but I've put them in a few of my amps just using one node and they work fine, seem quiet, and don't oscillate. Using one node yields higher plate voltages on the CF but that hasn't been a problem.

I did that in a Fender type build with a doghouse too. I'm using a single node, and the 100K across pins 1-6 trick like Marshall CF's. It sounds fine (I was prepared to put another node inside on terminals but I don't thing it's necessary).

I've also decided that I like a parallel loop better. I'm using the CF driver for both sides (dry and effect) then the mixer is a feedback amp with no ground reference. It's more transparent than the series loop. I also noticed in GP this week, that Fuchs uses a series/parallel switch on the loop. Sounds interesting. (Glowing review of the Triple Drive 150 BTW)

[Fischerman]

I'm using a single node, and the 100K across pins 1-6 trick like Marshall CF's. It sounds fine (I was prepared to put another node inside on terminals but I don't thing it's necessary).

Yea, that's how I always do it. I actually DO have an extra node that I shoehorned in my HRM...but the extra node is just to feed the actual FX loop node. So I suppose I could make it 'right' (but without the 'extra' node) but I don't see the need...it works and sounds good.

I also used dogears suggestion to increase the 1k5 and 10k resistors to 1k8 and ~27k...I went with 1k8 and 33k (all I had) which gives me about 45vdc on the cathode with 325vdc on the plate. My Return plate is at 225vdc.

[Tonegeek]

I don't really have the room either for the third node, and really my loop works fine now, but some time ago (and again here) Scott had suggested pulling it near the screen supply rather than later in the string to minimize the impact on the rest of the pre-amp supply. I have been feeling the need to build another amp so I am just planning that now and want to implement all that I have learned up to this point which includes the loop supply. I may at least move my 2 node loop supply in my existing rig next time I have the chassis out.

About a parallel loop, that is something I would like to try sometime. I have not had much luck with commercial built parallel loops (mostly due to phaseing problems), but I am encouraged to roll my own based on comments here. One thing I found out about my FX unit is that it inverts the phase, so that in itself seems problematic. Now that I think of it, I was probably not panning the FX unit to only emit the wet signal in which case phase is less important. AH HA! Wonder if I can mod my existing dlator for parallel without more tubes?

[Bob-I]

Now that I think of it, I was probably not panning the FX unit to only emit the wet signal in which case phase is less important.

Yes, that's critical. You can get phase problems if you mix both in the device and on the loop.

AH HA! Wonder if I can mod my existing dlator for parallel without more tubes?

You should be able to. Here's the schematic I used. Notice there's NO grid load on the mixer/gain stage. The feedback loop creates a virtual ground. Very strange at first when you measure the cathode and get 0 volts there, you have to measure against the virtual ground.

[groovtubin]

Now that I think of it, I was probably not panning the FX unit to only emit the wet signal in which case phase is less important.

Yes, that's critical. You can get phase problems if you mix both in the device and on the loop.

AH HA! Wonder if I can mod my existing dlator for parallel without more tubes?

You should be able to. Here's the schematic I used. Notice there's NO grid load on the mixer/gain stage. The feedback loop creates a virtual ground. Very strange at first when you measure the cathode and get 0 volts there, you have to measure against the virtual ground.

Bob, how in the world did you dream that one up! lol! AWESOME!

jim

[Tonegeek]

Notice there's NO grid load on the mixer/gain stage. The feedback loop creates a virtual ground. Very strange at first when you measure the cathode and get 0 volts there, you have to measure against the virtual ground.

Think i read something about this on another thread discussing the TUT loop. Still don't understand the virtual ground concept. It hurts my head to think about it