|'); } // End -->|
I'm about to build a 'Karna' style amp with DC filaments on the DHT driver . I notice Lynn Olson uses an unspecified gapped PP interstage but I'll be using the LL1635PP which can take 1mA of DC offset . I can either use 3A/110A at 8mA (Ra 5k mu 12) or preferably class A connected type 46 . I'm well aware that the 1635 requires tightly balanced valves but what's the consensus regarding the above circuit ? Fils are supplied by DC , the switch and resistors between B+ and the LL1635 are there to provide a bias tap . Is this sort of scheme worth the hassle or is it best just to reverse one of the filaments to adjust bias ?
each 46 filament has it's own DC supply as per the above schemo . what are the disadvantages of this circuit ?
"or is it best just to reverse one of the filaments to adjust bias ?"
How does doing that adjust the bias?
You need to adjust the *relative bias* on the two tubes to get an accurately identical current in each half of the stage. I don't see any provision to do this in your circuit.
'How does doing that adjust the bias?'
It uses DC filaments . One side of the filament is more positive than the other , reverse the filament connection and the operating point changes
OK, but a bit like a 2 position gas pedal on your car - not exactly what you would call a fine adjustment...
To get the accurate DC balance you want (and need) - an adjustable bias balance will be needed. Not just for the initial setting but to allow for tube drift across time. My DHT diff poweramps have a top panel metter and the 10 turn trim pot. needs a tweak about once a month - after the first month - during the first month of new tubes it needs a tweak every day...
Agreed on the multiturn pot .
'To get the accurate DC balance you want (and need) - an adjustable bias balance will be needed'
Thats on the original schematic , slightly crude I know . I'm well aware about tight biasing , every amp I've built since 1999 has had a centre-reading balance meter . Just wondering : Do you derive the variable bias of your DPA amps from a negative supply or from a pot in the cathode bias circuit ?
A couple of things I would change. Since you're using DC heating, it makes no sense in adding hum bucking pots on each side. I would replace the pots with fixed resistors - a pair of 10 ohms on each side would be okay and then take the signal from the center of those.
Next for balance control I would add just one pot - the wiper would go to the center resistor that heads to ground. Each other end would go to the common connection of those 10 ohm fixed resistors on each 46. I don't know what value would be best for that one pot, but it shouldn't need to be too big.
But you actually meant the B+ balance pot wiper should be hooked up to B+, not to the "ground" ??
Hi Kurt ,
Those aren't humbucking pots . The filaments are not shared , each pot provides a small amount of bias variation . I know it's a bit crude . The other option is like the biasing on the Williamson output stage , but maybe the best solution is a 5mA LL1635-PP instead .
This balancing is going to be a PITA. Better to make it less critical. The downside is that you'll lose a lot of induction so rules out high Rp drivers. I'm going to replace my IT's with custom ones that allow for 10mA DC in the near future.
I believe the filament bias idea would work better with CCS fils and a variable series resistor . Change in parameters due to ageing is the main concern . I'm already considering the 5mA offset version . The 1635PP would be ideal with the smaller 3A/110A DHT but I'd prefer to send my type 46 into action . Whatever I'll use , the valves will be burnt in and matched on the Avo , then matched in situ in the amp . The output stage is next on the hitlist , I only have a 5mA offset margin with the c-core output transformer .
Is there any reason not to use a CCS under the diff pair? Always worked for me. You can also wire the filamants of the two valves in parallel with a current source as well. You have a tester so should be able to sort some matching valves - I have quite a few 46 if you need to do any matching. Andy
Hi Andy ,
I may take you up on your offer , I have a dozen type 46 . Got any globes ? No CCS anywhere in this amp apart from in the B+ feed to the driver (2 x 105V gas regs not shown in the diagram) . The whole point is that the LL1635 interstage can only tolerate 1mA of unbalanced DC . Even tightly matched valves will drift with age so there needs to be some form of adjustment . In the diagram each filament has it's own DC supply , the pots provide a small amount of balancing . If the pots are removed then the '-' of each filament would be joined together , reversing the 'phase' of one filament will allow a change of 2.5V of bias . Alternatively the 3A/110A running at lower current can be used without any balancing , but the output stage will no longer be able to be driven hard into A2 . Ra for the 3A/110A is also higher than I'd like (5k) compared to the 46 (2.3k) . Sowter 8920 were delivered this morning and I've had an offer on the LL1635 . Chassis bashing will commence early nest week so I need to pull my finger out ;)
I've got ST 46s apart from a few globes I want to hang onto. I'm waiting to build something very similar - I have LL1635PP and LL1660/5 to put into a 3 stage amp. If you want to try out the LL1635 for measurement purposes I shouldn't need it for the next month or so. The amp I'm currently listening to has the LL1621/6 which seems to work OK - it's a close cousin of the LL1635. I've been using constant current sinks with some success, and the single filament supply goes to both of the diff pair tubes in parallel. All I can say is that it certainly works that way. Also very silent, which is nice. My present amp goes 3A5 to 1H4 to 2a3 - not ideal but I ran out of space so used smaller tubes. Sounds very detailed, and since it's entirely DHT, so it should! Andy
Hi Andy ,
I'm not too sure about the 1635 now , I'll probably go for the 1635 5mA , the 46 has nice low Ra so should be ok with the lower inductance of the 5mA version . How many type 46 did you buy , 100 ? ;) Your amp sounds nice but why not split the filaments ? I'll be doing my amp 4 box , 2 psu's on sand box plinths , 2 power sections on top . It'll be all DHT too , no IDHT as concertina valve this time .
How many type 46 did you buy , 100 ? ;)> > >
More like 50, all in bits and pieces from various sources.
Your amp sounds nice but why not split the filaments?> > > >
No room! It's all on one chassis. That's four filamant supplies as it is. Next amp goes on two mono chassis - double the room. I'm trying to sort out some PCBs for the filament supplies - that'll make things easier. With DHTs you need to prepare quite some stuff for the filament supplies you're going to need!!! That's one of the secrets of the sound, as you know.
I'll be doing my amp 4 box , 2 psu's on sand box plinths , 2 power sections on top . It'll be all DHT too , no IDHT as concertina valve this time. > > >
This post is made possible by the generous support of people like you and our sponsors: