Am currently planning to build a PP 6B4G amp similar to Ralph Power's unit. I would like to drive the finals with an ECC99 and have read through the threads. I have two questions;
1. I have a pair of Magnequest Trans from an ST-70 clone. They have 4.3 K primary impedance. Would that match up well with the 6B4G?
2. What would be your suggested configuration for the drivers looking into a Lundhal LL1630 interstage Trans?
Sorry for the basic questions but my tube theory is just good enough to prevent me from killing myself or burning down the house!
Would appreciate your input.
Regards;
Pete
1. I have a pair of Magnequest Trans from an ST-70 clone. They have 4.3 K primary impedance. Would that match up well with the 6B4G?
2. What would be your suggested configuration for the drivers looking into a Lundhal LL1630 interstage Trans?
Sorry for the basic questions but my tube theory is just good enough to prevent me from killing myself or burning down the house!
Would appreciate your input.
Regards;
Pete
Thanks for your responses; I believe that I would only require 6-8 Watts since I am using speakers whose sensitivity is 94db and only above 125hz. I had thought to keep the finals in class 'A' if possible thereby only asking for perhaps a -45/+45 swing from the ECC99's.
I am currently using the output of a CD player to drive the system and had not planned on building a linestage with a balanced output.
Pete
I am currently using the output of a CD player to drive the system and had not planned on building a linestage with a balanced output.
Pete
LTCP?
Hi,
Hmmm....Vaguely remember that one.
Was this a cascode of 6922s as LTP with a CCS or a negative Vref in the tail?
I think this would have tremendous gain (1000+) and very low noise as well using the 6922.
Alternatives could be the 6SN7 (µ 420) or a 6SL7(µ 5000).
From this a ECC99 could be driven which in turn would rattle the grids of the 6B4Gs.
Cheers,
Hi,
Hmmm....Vaguely remember that one.
Was this a cascode of 6922s as LTP with a CCS or a negative Vref in the tail?
I think this would have tremendous gain (1000+) and very low noise as well using the 6922.
Alternatives could be the 6SN7 (µ 420) or a 6SL7(µ 5000).
From this a ECC99 could be driven which in turn would rattle the grids of the 6B4Gs.
Cheers,
Yes, that's it. 50K plate load, 400V B+, 10ma CCS in the cathode circuit. Gain is a bit lower than your guess, about 300-400, IIRC. I don't know if you'd even need the extra ECC99 driver tube- the 6922/6DJ8 family can pump pretty well into triode grids.
You'd implement it with CCa, if I know you.
You'd implement it with CCa, if I know you.
Hi,
Nah...You probably won't even need the extra stage.
Your gain "guesstimate" seems rather low considering the humble SN7 can make it to 420 (checked the math on that one).
This is a formidable circuit idea that's both simple and powerful, I'm starting to like it already.
The few I've left are sitting in my headamp....
With 24 VDC on their roof and 6V on the heaters they're going to last forever....
Not a bad thing either, considering the asking price.
BTW, the LTPC was first implemented_that I'm aware of_ and publicised by L.B. Hedge in a 1956 Wireless World issue page 283. "Cascode A.F. Amplifier".
No need to tell J.C.
Nah...You probably won't even need the extra stage.
Your gain "guesstimate" seems rather low considering the humble SN7 can make it to 420 (checked the math on that one).
This is a formidable circuit idea that's both simple and powerful, I'm starting to like it already.
The few I've left are sitting in my headamp....
With 24 VDC on their roof and 6V on the heaters they're going to last forever....
Not a bad thing either, considering the asking price.
BTW, the LTPC was first implemented_that I'm aware of_ and publicised by L.B. Hedge in a 1956 Wireless World issue page 283. "Cascode A.F. Amplifier".
No need to tell J.C.
Agreed, mu is very high for a cascode, but it suffers in the same way as a pentode from high output resistance, and that drops its gain. I recently screwed a gain of 200 out of a 6SN7/FET differential pair cascode. (Predicted gain = 235)
Edit: Should have said that the gain of 200 was to one anode.
Edit: Should have said that the gain of 200 was to one anode.
Hi,
O.K. I think I see where the confusion stems from.
My mistake, I quoted figures for the cascode. Not the gain at the anode node.
The circuit I have here takes the signal off the top anodes, a trimpot allows for balance adjudment between the two top triodes' anode resistors.
The top pair grids are tied together and fed from the same rail as the anodes but at lower potential.
NFB can be applied to the grid of the second lower triode which tied to ground through a gridleak resistor. (Grounded grid inverter part of the LTP)
In the circuit the output tubes are 1625s, a variant of the 807.
The UTC LS-55 OPT is switchable for UL, penthode and triode mode.
Cheers,
P.S. Think I'll go to bed with Joe tonight to refresh the glasshard memory.
O.K. I think I see where the confusion stems from.
My mistake, I quoted figures for the cascode. Not the gain at the anode node.
The circuit I have here takes the signal off the top anodes, a trimpot allows for balance adjudment between the two top triodes' anode resistors.
The top pair grids are tied together and fed from the same rail as the anodes but at lower potential.
NFB can be applied to the grid of the second lower triode which tied to ground through a gridleak resistor. (Grounded grid inverter part of the LTP)
In the circuit the output tubes are 1625s, a variant of the 807.
The UTC LS-55 OPT is switchable for UL, penthode and triode mode.
Cheers,
P.S. Think I'll go to bed with Joe tonight to refresh the glasshard memory.
Hi,
I had a feeling that would be the case....
However, I reread Joe C.s in G.A. last night...actually early this morning, and although he credits the same article by Hedge as I mentioned earlier, he doesn't give us any new formulas for gain calculation of his circuit.
So, using what I have on hand for cascode:
µ' = µ (µ +1)
I arrive at :
µ' = 33 (33+1)
µ' = 33*34
µ = 1122
Which is close to my 1000+ guess earlier.
Now, if this is correct and taken into account the fact that we have penthode like behaviour ( as EC8010 correctly stated), it remains to be seen if the cct is capable of driving a PP of 6B4Gs adequately...
Hence my suggestion earlier to use the ECC99 as an extra driver....
Which kind of defeats the use of the LTPC in the first place, so maybe using that ECC99 iso the ubiquitousD) 6KN8 as LTPC may solve the drive problem...
Cheers,
I had a feeling that would be the case....
However, I reread Joe C.s in G.A. last night...actually early this morning, and although he credits the same article by Hedge as I mentioned earlier, he doesn't give us any new formulas for gain calculation of his circuit.
So, using what I have on hand for cascode:
µ' = µ (µ +1)
I arrive at :
µ' = 33 (33+1)
µ' = 33*34
µ = 1122
Which is close to my 1000+ guess earlier.
Now, if this is correct and taken into account the fact that we have penthode like behaviour ( as EC8010 correctly stated), it remains to be seen if the cct is capable of driving a PP of 6B4Gs adequately...
Hence my suggestion earlier to use the ECC99 as an extra driver....
Which kind of defeats the use of the LTPC in the first place, so maybe using that ECC99 iso the ubiquitous
D) 6KN8 as LTPC may solve the drive problem...Cheers,
Hi,
Can't disagree with that....From past experience with stacked triode a la SRPP using, roughly, µ/2 iso of µ in the formula may be closer to reality and even that is too optimistic IMHO.
Now if we could arrive at a formula that would be accurate at gain calculation and at the same time not muck up other parameters in the process, we would have a winner on our hands.
BTW, are you still in touch with J.C.?
Maybe he could shed some light on this...Just wondering though, I know you're busy chasing young Jimmy around.
Or maybe EC8010 could enlighten my clouded mind...
Cheers,
Can't disagree with that....From past experience with stacked triode a la SRPP using, roughly, µ/2 iso of µ in the formula may be closer to reality and even that is too optimistic IMHO.
Now if we could arrive at a formula that would be accurate at gain calculation and at the same time not muck up other parameters in the process, we would have a winner on our hands.
BTW, are you still in touch with J.C.?
Maybe he could shed some light on this...Just wondering though, I know you're busy chasing young Jimmy around.
Or maybe EC8010 could enlighten my clouded mind...
Cheers,
Just because it's been made into a differential pair doesn't change the gain, it's just that you see the output voltage between the two anodes, so you can still use a cascode gain equation. The easiest to use is simply;
Av = gmRL
Where:
gm = mutual conductance of lower valve
RL = load resistance at the upper valve
This equation tends to give a slightly optimistic value for gain, but posting more complex equations is hard.
Av = gmRL
Where:
gm = mutual conductance of lower valve
RL = load resistance at the upper valve
This equation tends to give a slightly optimistic value for gain, but posting more complex equations is hard.
Wayne; thanks for the data, 6+ Watts should be just fine.
Frank; Sy; your comments kind of leave me in the dust, as I said my tube theory is old and rusty.
What does LTPC stand for?
I do not have the article from Glass Audio but it sounds like you are describing a Cascoded Differential Amplifier yes?
The Power's amp got around 8 Watts with a single stage 5842/417A coupled to the 6B4G's by the Lundhal 1635 Trans, would that same result be possible with the ECC99?
Sorry for the dumb questions,
Pete
Frank; Sy; your comments kind of leave me in the dust, as I said my tube theory is old and rusty.
What does LTPC stand for?
I do not have the article from Glass Audio but it sounds like you are describing a Cascoded Differential Amplifier yes?
The Power's amp got around 8 Watts with a single stage 5842/417A coupled to the 6B4G's by the Lundhal 1635 Trans, would that same result be possible with the ECC99?
Sorry for the dumb questions,
Pete
Hi,
Long Tailed Pair Cascode.
You guessed it right.
The available power will depend on how the output tubes are run.
The gain of the amp will need to come from the input tube + (in this case the IS xformer), µ of the 5842 is more than twice that of a ECC99 , the latter being comparable to, say, a super 6SN7.
To my mind, I think SY's as well, the IS xformer isn't really needed.
After all there's no iron like no iron, right?
Which is why we took this philosophical leap into the LTPC.
Cheers,
Long Tailed Pair Cascode.
You guessed it right.
The available power will depend on how the output tubes are run.
The gain of the amp will need to come from the input tube + (in this case the IS xformer), µ of the 5842 is more than twice that of a ECC99 , the latter being comparable to, say, a super 6SN7.
To my mind, I think SY's as well, the IS xformer isn't really needed.
After all there's no iron like no iron, right?
Which is why we took this philosophical leap into the LTPC.
Cheers,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.