EF80, 86 are going to have a hard time supporting the current without making the screens glow orange. ECL82 is a bit overkill, but not enough overkill where it won’t work.
ECL82 curves look bad around 100V. I would not use that. From the tubes you got the EF80 is the best bet. I woukd get a NOS EF184 from e-bay, they are quite cheap.
The output resistance of the current source only has to be much greater than the reciprocal of the transconductance of the second valve of the long-tailed pair, say 99 times greater if you allow a 1 % current loss. I think you can meet that requirement very well with a triode with non-decoupled cathode resistor (or with a pentode).
I once used a Hammond 193H 5H 200mA choke as a 'CCS' for a pair of 6C45pi triodes for a cathode coupled phase splitter. The 193H has 65 Ohms DCR, the "CCS" burden voltage was really nice and low. Negative supply is not needed, just connect the choke to ground, and the other end to the parallel cathodes.
The high transconductance of the 6C45pi tubes easily overcame the 5H inductance, that circuit worked reasonably well all the way down to 20Hz.
The high transconductance of the 6C45pi tubes easily overcame the 5H inductance, that circuit worked reasonably well all the way down to 20Hz.
I think its better to get a proto going before attempting the CCS. I have 4 or 5 days window every 20 odd days for DIY.
What is the opinion on the ECC88, EL34 pp schematic posted in #14? I have some used Russian equivalents with me and output iron 4600R/8R.
What is the opinion on the ECC88, EL34 pp schematic posted in #14? I have some used Russian equivalents with me and output iron 4600R/8R.
1. No...the long tail pair is made up of two halves, the first half is common cathode, but is cathode couple to the second half which is a common grid, the first half is inverting, the second half is non inverting, so that two plates are resistor loaded and produced out of phase signals..Have a set of quick questions on the DC Long Tail pair splitter.
Schematic cut/pasted from https://www.angelfire.com/electronic/funwithtubes/index.html
- Just where the splitting happens? Across the 1M resistor?
- What is the function of that 0.22uf cap from grid to the ground? Would increasing it to 1uf make any impact on overall frequency spectrum? How critical is the quality of the cap used here? Cheap cap can do?
- Anode resistors need to be matched? In this instance one is 33k while the other is 39k. Purpose?
View attachment 1239396
2. the 0.22 effectively grounded the grid of the second half at ac, the 1Meg dc biased the grid of the second half, so that both grids are biased at the same dc voltage..
3. not necessarily, but some do load the halves with different value resistors, the tail current ensured that the output are balanced all the time...
No. To get things moving I intend to use a normal resistor (value?) and when time permits put a current source.Post # 14:
The 5687 plate impedance will be approximately 25k Ohms.
Compare that to the two E88CC cathode impedances it powers.
That tells you about the effectiveness of the 5687 "CCS".
I already have a donor DIY amp that uses a design similar to my first post. Originally it used ECC83/ECC82 but now I wish to rebuild with a design that can hopefully use many 6N1P tubes that I have.
I would likely used 6N1P in place of 5687 in the first socket as well.
Tolerance of valve parameters means it's likely that output signal balance, and dc idle conditions, will noticeably vary even for NOS tubes, and certainly for aged tubes. Given that you may then be driving a driver and/or output stage, which has its own imbalance, may make it relevant to include some balance adjustment - especially if you aim to minimise distortion by adjustment using a distortion meter or spectrum analyser. A common location for a balance trimpot is with the PI stage anodes to B+, which offsets the need to initially rely on 'theoretically' derived anode resistor values,
VOB,
OK. Resistor in the coupled cathodes circuit.
Remember to make the plate resistor of the tube with the un-driven grid, more resistance than the plate resistor of the tube with the driven grid.
Those different resistor values are approximations to make the two plate signal swings the same amplitude.
The ratio is based on the resistance of the common cathode resistor, versus the operating cathode impedance.
Tektronix tube scopes used to use -150V supplies, so that the common cathode grid resistor's large value, made it possible to use equal load resistors in the plates (and sometimes those plate load resistors were in series with inductors for high frequency peaking).
Most ideas about tubes are borrowed, stolen, modified, or all three.
Be thankful for Edison, Fleming, DeForest, Armstrong, RCA, Bell Labs, Western Electric, Raytheon, HP, Tektronix, Collins, Radio Amateurs, Hedy [hopper] Lamarr, Shannon, and way too many others to mention . . . they made it possible to have what we have.
OK. Resistor in the coupled cathodes circuit.
Remember to make the plate resistor of the tube with the un-driven grid, more resistance than the plate resistor of the tube with the driven grid.
Those different resistor values are approximations to make the two plate signal swings the same amplitude.
The ratio is based on the resistance of the common cathode resistor, versus the operating cathode impedance.
Tektronix tube scopes used to use -150V supplies, so that the common cathode grid resistor's large value, made it possible to use equal load resistors in the plates (and sometimes those plate load resistors were in series with inductors for high frequency peaking).
Most ideas about tubes are borrowed, stolen, modified, or all three.
Be thankful for Edison, Fleming, DeForest, Armstrong, RCA, Bell Labs, Western Electric, Raytheon, HP, Tektronix, Collins, Radio Amateurs, Hedy [hopper] Lamarr, Shannon, and way too many others to mention . . . they made it possible to have what we have.
Last edited:
trobbins,
Correct. No calculated theoretical values are perfect on a real circuit; parts are not perfect.
Check your average resistor, with parasitic C and L.
Matched tubes, can be very close, but not exact, especially at large signal values.
I used to measure the series inductance of coupling caps on the VNA. Get a load of that impedance null!
You need a spectrum analyzer to check the 2nd harmonic distortion, or you are just guessing that the balance pot is nulling the 2nd harmonic distortion. (It often changes the 3rd harmonic distortion as well, so a simple THD analyzer can not guarantee the lowest THD is also the 2nd harmonic null).
If you check the spectrum at the amplifier output, you are not sure which un-balanced circuit you are actually compensating for; perhaps you are making the sum of the phase splitter and output stage cancel the Total 2nd harmonic distortion.
Do you check the balance With global negative feedback enabled, or with global negative feedback disabled?
Depends on what you want, balance of the whole amplifier, or balance of the phase inverter circuit>
Just my experience with balance pots.
Your Mileage May Vary.
Correct. No calculated theoretical values are perfect on a real circuit; parts are not perfect.
Check your average resistor, with parasitic C and L.
Matched tubes, can be very close, but not exact, especially at large signal values.
I used to measure the series inductance of coupling caps on the VNA. Get a load of that impedance null!
You need a spectrum analyzer to check the 2nd harmonic distortion, or you are just guessing that the balance pot is nulling the 2nd harmonic distortion. (It often changes the 3rd harmonic distortion as well, so a simple THD analyzer can not guarantee the lowest THD is also the 2nd harmonic null).
If you check the spectrum at the amplifier output, you are not sure which un-balanced circuit you are actually compensating for; perhaps you are making the sum of the phase splitter and output stage cancel the Total 2nd harmonic distortion.
Do you check the balance With global negative feedback enabled, or with global negative feedback disabled?
Depends on what you want, balance of the whole amplifier, or balance of the phase inverter circuit>
Just my experience with balance pots.
Your Mileage May Vary.
Last edited:
Its getting interesting now.VOB,
OK. Resistor in the coupled cathodes circuit.
Remember to make the plate resistor of the tube with the un-driven grid, more resistance than the plate resistor of the tube with the driven grid.
How do we go about calculating the values of plate and cathode resistors for this LTP.
What play does they have the with preceding stage and the finals which in our case in EL34 in PP.
I am planning to use type ECC88 tubes for input and LTP.
A note about the circuit in post #1: the circled 0u22F capacitor would ideally go to the "top" of the first stage's anode load resistor rather than to signal ground. Gives some PS noise immunity and removes residual imperfection of the bypass capacitor, by taking advantage of the LTP's natural differential input.
All good fortune,
Chris
All good fortune,
Chris
- Home
- Amplifiers
- Tubes / Valves
- DC Long Tail pair splitter question