I am designing an AB2 amp with 807-tubes. At the desired output power I need to drive the grids toward +10V. At this voltage the grids form a voltagedependant resistor, which can be calculated at approx 1100Ohms at +5V and approx 700Ohms at +10V. At +10V the cathode folower should be able to deliver 15mA to the grid plus 5mA to a cascode current sink in the Cathode. So I need a triode with high µ and low ra for a cathode folower to meet these demands.
I have been looking at the triodes in the attached image-file.
At the extreme right columns the attenuation due to the output resistance is calculated. ECC83 is just shown as reference. ECC88 has the lowest outputresistance, but it does not meet the requirements on current delivery at 20mA. 6N6P seem to be the best choise of theese triodes.
The question is if there are better triodes for the job out there that I have overlooked?
Regards spuncut
I have been looking at the triodes in the attached image-file.
At the extreme right columns the attenuation due to the output resistance is calculated. ECC83 is just shown as reference. ECC88 has the lowest outputresistance, but it does not meet the requirements on current delivery at 20mA. 6N6P seem to be the best choise of theese triodes.
The question is if there are better triodes for the job out there that I have overlooked?
Regards spuncut
Attachments
Last edited:
the best triode for the job is a mosfet.
Either Tubelab's power drive or Pete Millett's Class A2 buffer
Either Tubelab's power drive or Pete Millett's Class A2 buffer
You should be looking at tubes more like the 6BL7, 6BX7, 6S4A, 6080, 6AS7G, etc.
You haven't said a whole ton about your amp, but your B+ rail is going to be quite high I'm sure. A tube like the 6AS7G could do double duty with one triode working as a cathode follower and the other triode acting as a series regulator to knock down the B+ appearing at the plate of the cathode follower triode.
You haven't said a whole ton about your amp, but your B+ rail is going to be quite high I'm sure. A tube like the 6AS7G could do double duty with one triode working as a cathode follower and the other triode acting as a series regulator to knock down the B+ appearing at the plate of the cathode follower triode.
More about the amp
Actually I only need +5V on the grid of 807, but in order to have some headroom I design for +10V.
See attached file for loadline.
Z_T is the OT-impedance per anode.
V_Amin, Cyan line, is the lowest anodevoltage.
V_A, Chocking pink line is the B+.
K_bias, Green slider is the bias current divided by Max anode dissipation (P_Amax, red)
When the G1-voltage is pushed to +10V the anode voltage (V_Amin) swings down to 67V (and the screen current goes nuts).
Regards spuncut
Actually I only need +5V on the grid of 807, but in order to have some headroom I design for +10V.
See attached file for loadline.
Z_T is the OT-impedance per anode.
V_Amin, Cyan line, is the lowest anodevoltage.
V_A, Chocking pink line is the B+.
K_bias, Green slider is the bias current divided by Max anode dissipation (P_Amax, red)
When the G1-voltage is pushed to +10V the anode voltage (V_Amin) swings down to 67V (and the screen current goes nuts).
Regards spuncut
Attachments
Last edited:
Peak cathode current is worth looking at for these tubes. I think either the BL7 or the BX7 has a peak cathode current well over 100mA.
Why high Mu? It is a cathode follower.
What you want is high Gm and moderately higher current than most small-signal triodes. Mu matters only as a second-order trend.
Also remember that in a push-pull amp, only one side can be drawing grid current. It is advantageous to link both sides, say with a center-tap choke, so both drivers can pull each peak.
To PRR:
Yes it's a triode for a low output resistance cathode folower iam looking for. Sorry for not making it clear in the initial post, English is not my first language.
Due to the formula:
CFRout = ra/(µ+1)
...ra should be low and µ should be high in order to obtain a low output resistance.
About the Gm, please educate me.
Utilizing a centertapped choke sounds interesting can you point me to a example schematic?
To audiowize:
About 30mA will be adequate, the screen will sag before the G1 current becomes excessive.
If you look at the table in post #4 the 6N6P still seem to be the best choice.
Regards spuncut
Yes it's a triode for a low output resistance cathode folower iam looking for. Sorry for not making it clear in the initial post, English is not my first language.
Due to the formula:
CFRout = ra/(µ+1)
...ra should be low and µ should be high in order to obtain a low output resistance.
About the Gm, please educate me.
Utilizing a centertapped choke sounds interesting can you point me to a example schematic?
To audiowize:
About 30mA will be adequate, the screen will sag before the G1 current becomes excessive.
If you look at the table in post #4 the 6N6P still seem to be the best choice.
Regards spuncut
Last edited:
Don't neglect that CFRout=RA/(u+1) is almost the same as CFRout=1/gm since mu=gm*rp, mu/rp=gm, so rp/mu=gm^(-1). Highest gm is desirable, but ultimately that tends to come with limited plate voltage. Triode strapped EL34 and EL84 could certainly be in the running as well.
Thanks for putting me right on the tubemath. 🙂
EL34 and EL84 for driving a 807 seems wildly overkill to me.
Regards spuncut
Agree, however if you must use a tube, 5687 will make a decent CF, and suggest giving them their own heater supply to avoid breaking H-K voltage limits
FYI my P-P amps are triode-strapped 6L6 with MOSFET followers DC coupled to o/p grids, they work very well and I believe that the MOSFETs are sonically transparent
Radiotron, do you use Pete’s solutuion with a MOSFET and a CCS. Can you give us a diagram? I’m working on something similar for driving KT150’s.
Regards, Gerrit
Regards, Gerrit
In the 6L6 P-P, no, there's a 15k 2W metal oxide from source of MOSFET to B-
And used Tubelab's idea for power supply to the follower with a 25VA 115-0-115 isolating transformer, with a simple C-R-C filter
I don't have a schemo handy as I'm about to run out the door to work, will post something later if nobody else can help
And used Tubelab's idea for power supply to the follower with a 25VA 115-0-115 isolating transformer, with a simple C-R-C filter
I don't have a schemo handy as I'm about to run out the door to work, will post something later if nobody else can help
I didn't review your load lines all that carefully, but I would assume you are going for the ~80W operating parameters in the datasheet.
Gerrit, these pictures from notebook might be of use, I don't have an up to date schematic of the complete amp. The power supply is not critical so long as it is quiet, a C-R-C filter simmed OK in PSUD2 and works fine in practice.
The finals are 6L6GA with 350V HT, adjusted to idle at 40mA each, looking into approx 6k plate to plate, about 13-15 watts out, enough for my 96dB/1W speakers.
The finals are 6L6GA with 350V HT, adjusted to idle at 40mA each, looking into approx 6k plate to plate, about 13-15 watts out, enough for my 96dB/1W speakers.
Attachments
I recently prototyped a AB2 807 amp and tried various valves but a 12BH7 CF did the job well supplied from a -80v rail. I bought two 6BL7's for a future amp to do this job but 12BH7's and 6SN7's are used in several 807 AB2 designs.
If you look at the STC 807 datasheet there's an AB2 design in there to give you a start. If you also read the datasheet it emphasis's the need for a well regulated screen grid supply. I tried a AB2 UL design but this killed an 807, not surprising as there was over 400v on the screen grids.
Andy.
If you look at the STC 807 datasheet there's an AB2 design in there to give you a start. If you also read the datasheet it emphasis's the need for a well regulated screen grid supply. I tried a AB2 UL design but this killed an 807, not surprising as there was over 400v on the screen grids.
Andy.
Most triodes will work when NFB is applied, that is, if they can provide the current. But openloop data provided by design are, in most cases, better than closedloop-data obtained by NFB. So I would like to reduce the distortion in the B2-range as much as possible. The distortion occurs when the CF's output resistance is driving the non-linear inputresistance of the 807's G1. Hence the lower outputresistance of the cf-driver the lower the distortion (see the table in post #4 where different triodes are compared).
The use of mosfets are effective to obtain low driver resistance, but I'll stick with cf-drivers for now.
Provided that my calculations are correct the best suited triode for the driver is 6N6P, but I'm still looking for better suggestions.
Regards spuncut
The use of mosfets are effective to obtain low driver resistance, but I'll stick with cf-drivers for now.
Provided that my calculations are correct the best suited triode for the driver is 6N6P, but I'm still looking for better suggestions.
Regards spuncut
Last edited:
This version appeared many years ago, used as a modulator in ham transmitters. So we know it works.🙂
