Some time ago I purchased a box of 6P43P pentodes.
At the time there was little data and no plate curves available on these tubes and I was not too sure what use they could be put to.
As they were under a dollar each I figure what the heck.
Running the 6P43P thorough my uTracer3+ with 200V on the screen it provided over 200mA of plate current at -6V on the grid.
That is way ahead of a 6BQ5 in terms of delivering plate current and it seems these could drive lower impedance loads than the suggested 6BQ5 8K PP load.
The plate curves also looked pretty linear with very little kinking at low plate currents.
A data sheet value of 12 watts plate dissipation and 2 watts screen dissipation also looked good.
Va maximum is shown on the data sheet as 300V same as the 6BQ5.
The 6P43P bias point for 35mA plate current at a Vs of 250V is Vg1 of -26 volts.
The 6BQ5 under the same Vs=250V, 35mA plate current requires about Vg1 of -9.5 volts.
The 6P43P gm is stated as 7.5mA/V and the 6BQ5 at 11.3mA/V
So it would seem a 6P43P is going to require 8.7dB more drive voltage and a input stage gain increase over a 6BQ5 design.
To keep the design small and simple I decided to limit the front end to a single stage.
I also wanted to use significant amounts of local feedback to reduce the total global feedback required around the output transformer to improve stability.
That meant I would need a lot of gain in that single gain stage. I figured I would need over 44dB of input gain.
A pentode stage can provide a lot of gain however the output impedance tends to be higher limiting bandwidth.
A triode with this type of u did not seem possible.
Cascode stages can provide a lot of gain. One has to be careful of the output impedance or bandwidth will be limited.
In a Cascode stage total gain is largely a function of S in mA/V of the lower device divided into the top load impedance.
This means S, load impedance, bandwidth and gain can be traded as required.
The higher the tubes S the lower the load impedance can be and so the wider bandwidth for the same gain.
As I wanted a lot of gain and wide bandwidth I looked for the highest S tube I had on hand so the load impedance could be as low as possible.
The 6Z51P pentode operating in triode mode measures a S of 18~23mA/V with Va=85V and Ia=7.15mA.
Two tubes were arranged as a LTP. With a load impedance of about 23K ohms two 6Z51P operating in triode cascode mode yielded a balanced stage gain of 53dB and produced a balanced output of 48.7 RMS, 0.4% THD @1K with no feedback.
The low 23K total impedance provides wide bandwidth.
I used a pair of BJT for the top of the cascode.
This balanced input stage is coupled to a pair of 6P43P in PP using UL feedback for the screens with fixed bias.
The PP load impedance is 5K ohms and the plate supply is 330~342V.
The screens are feed from a regulated 275 volt supply and capacitor coupled to the UL taps on the transformer.
The amount of UL feedback was selected to result in 208V on the screen at the full power at lowest plate voltage point of the output swing to insure adequate plate current was available for full power output.
The bias is fixed and provided from a regulated power supply.
The result was surprisingly good for a pair of under 1 dollar tubes.
THD at 1 watt 1Khz was 0.0119%
THD at 1 watt 20Khz was 0.083%
THD at 1 watt 30hz was 0.0998%
THD at 19 watt 1Khz was 0.09%
THD at 19 watt 20Khz was 1.01%
THD at 19 watt 30hz was 0.599%
At the time there was little data and no plate curves available on these tubes and I was not too sure what use they could be put to.
As they were under a dollar each I figure what the heck.
Running the 6P43P thorough my uTracer3+ with 200V on the screen it provided over 200mA of plate current at -6V on the grid.
That is way ahead of a 6BQ5 in terms of delivering plate current and it seems these could drive lower impedance loads than the suggested 6BQ5 8K PP load.
The plate curves also looked pretty linear with very little kinking at low plate currents.
A data sheet value of 12 watts plate dissipation and 2 watts screen dissipation also looked good.
Va maximum is shown on the data sheet as 300V same as the 6BQ5.
The 6P43P bias point for 35mA plate current at a Vs of 250V is Vg1 of -26 volts.
The 6BQ5 under the same Vs=250V, 35mA plate current requires about Vg1 of -9.5 volts.
The 6P43P gm is stated as 7.5mA/V and the 6BQ5 at 11.3mA/V
So it would seem a 6P43P is going to require 8.7dB more drive voltage and a input stage gain increase over a 6BQ5 design.
To keep the design small and simple I decided to limit the front end to a single stage.
I also wanted to use significant amounts of local feedback to reduce the total global feedback required around the output transformer to improve stability.
That meant I would need a lot of gain in that single gain stage. I figured I would need over 44dB of input gain.
A pentode stage can provide a lot of gain however the output impedance tends to be higher limiting bandwidth.
A triode with this type of u did not seem possible.
Cascode stages can provide a lot of gain. One has to be careful of the output impedance or bandwidth will be limited.
In a Cascode stage total gain is largely a function of S in mA/V of the lower device divided into the top load impedance.
This means S, load impedance, bandwidth and gain can be traded as required.
The higher the tubes S the lower the load impedance can be and so the wider bandwidth for the same gain.
As I wanted a lot of gain and wide bandwidth I looked for the highest S tube I had on hand so the load impedance could be as low as possible.
The 6Z51P pentode operating in triode mode measures a S of 18~23mA/V with Va=85V and Ia=7.15mA.
Two tubes were arranged as a LTP. With a load impedance of about 23K ohms two 6Z51P operating in triode cascode mode yielded a balanced stage gain of 53dB and produced a balanced output of 48.7 RMS, 0.4% THD @1K with no feedback.
The low 23K total impedance provides wide bandwidth.
I used a pair of BJT for the top of the cascode.
This balanced input stage is coupled to a pair of 6P43P in PP using UL feedback for the screens with fixed bias.
The PP load impedance is 5K ohms and the plate supply is 330~342V.
The screens are feed from a regulated 275 volt supply and capacitor coupled to the UL taps on the transformer.
The amount of UL feedback was selected to result in 208V on the screen at the full power at lowest plate voltage point of the output swing to insure adequate plate current was available for full power output.
The bias is fixed and provided from a regulated power supply.
The result was surprisingly good for a pair of under 1 dollar tubes.
THD at 1 watt 1Khz was 0.0119%
THD at 1 watt 20Khz was 0.083%
THD at 1 watt 30hz was 0.0998%
THD at 19 watt 1Khz was 0.09%
THD at 19 watt 20Khz was 1.01%
THD at 19 watt 30hz was 0.599%
Attachments
-
6P43PE-PEN200-EX.txt
-
6P43PE-PEN250-EX.txt
-
6p43n-utracer3-Vs200V.gif -
1K @ 1WATT .0119THD.jpg -
6p43n 20khz 19watts thd.jpg -
6p43n 20khz 1watt thd.jpg -
1K @ 19Watt .09THD.jpg -
6p43n 30hz 1watt thd.jpg -
6p43n 30hz 19watts thd.jpg -
6p43n bottom.jpg -
6p43n top view.jpg -
6p43n transformer view.jpg -
3216AB10-SCH.PDF
-
3216AB10-PCB.gif -
20250408_120126.jpg -
20250408_120140.jpg -
20250409_172850.jpg -
20250408_120337.jpg
Last edited:
"Over 200mA plate current" ... for how long!
Here are the maximum ratings of the 6p43p-e
Technical specifications:
Maximum permissible parameters:
Operating time not less than 5000 hours.
Low anode voltage means low voltage swing.
Wishful thinking here I think.
Here are the maximum ratings of the 6p43p-e
Technical specifications:
- Heating voltage: 6.3 ± 0.7V
- Glow current: 625 ± 55mA
- Anode voltage: 185V
- Voltage on the second grid: 185V
- Resistance in the cathode circuit: 340Ohm
- Current in the anode circuit: 45 ± 9mA
- Current in the circuit of the second grid: not more than 2.7mA
- The steepness of the characteristic: 7.5mA / V
- Reverse grid current: not more than 1μA
Maximum permissible parameters:
- The highest voltage on the anode: 300V
- The highest voltage on the second grid: 250V
- Maximum power dissipated on the anode: 12W
- The highest power dissipated on the second grid: 2W
- The highest current in the cathode circuit: 75mA
- The highest voltage of the cathode-heater: 100V
- The greatest resistance in the circuit of the first grid: 2.2mOhm
Operating time not less than 5000 hours.
Low anode voltage means low voltage swing.
Wishful thinking here I think.
@JonSnell Electronic You are right in that the datasheet values would not allow such operation. However, those tubes are known to handle much higher ratings. @artosalo has a lot of experience with that tube:
I think you could be misinterpreting the data sheet values.
The 45mA current is the expected current for a typical tube under the parameter test conditions.
This number is not related to the tubes peak current capability.
The 75mA current rating is the maximum DC current that can be sustained.
Again this is not the peak current that can be supplied during a audio peak.
For example the 6BQ5 has a test parameter current of only 48mA and the maximum DC current is 65mA.
Endless designs over more that a 1/2 a century operate the 6BQ5's at peak currents far above these values.
Looking at 6BQ5 plate curves shows data for plate current up to 150mA and above with a 250V screen voltage.
In my design the DC current (bias current) is set to 35mA well below the tubes maximum rating of 75mA DC current.
The peak current that occurs briefly at the peaks of the sin wave a full power output is...
Assume a PP load 5k / 4 = 1250 ohms per tube during the class B portion at full power output.
So the peak current will be ( (19 watts / 1250 ohms)^0.5 ) * 1.414 = 174.33mA
Damage to the tubes cathode coating occurs if the space charge is depleted. Typical coated cathode tubes have a space charge that is many times the maximum cathode current.
I quote
"
Saturation
The current emitted by the cathode under normal conditions is far in excess of the current that passes to the plate. The emission of an oxide cathode is upwards of 0.5A/cm2. Even for a power tube, the ratio between the emitted current and the operating current is at least 10. The maximum plate current rating that appears in the specification for a tube is based mainly on considerations of heating, and in pulse operation can safely be exceeded (as it was in TV sweep circuits, for example).
"
From "https://www.john-a-harper.com/tubes201/"
If the space charge is 10 times the maximum DC current rating of 75mA then my peak current of 174mA would seem very safe operation.
Last edited:
The uTracer3+ (and other versions of this tracer) traces the tube under pulse condition. So, for most of the time the tube being traced/measured is not passing any current.
Thanks for your interest.
The schematic of the amplifier was included in my first post as the attached PDF file.
Unfortunately the forum can not provide a preview for attached PDF files so they are very easy to miss.
I did not post the power supply regulator schematics as there so many ways to provide regulated rails.
If you have interest I can post the power supply schematics as well.
Take care
The front end is a triode/MOSFET cascode. Effectively a pentode without the ability to swing its plate below screen voltage, but with no screen current. 6H6Pi and FQP1N60. Rides on a CCS to do phase splitting, and the LTP E-Linear rigged to the 28% taps of the OPT. Finals are 6P15P-EV running B+ of 300V thanks to an Aliexpress SMPS that does heater voltage as well. No measurements yet, power is adequate for open baffle EV Wolverine 8".
Douglas
Douglas
I would suggest there is a place for continuous measurements and pulse based measurements in tubes to probe various performance parameters.
For collecting parameters such as gm and bias voltage used in matching of power tubes I favor continuous measurements as the heating involved does cause small but meaningful parameter shifts with time in my experience.
The closer they are run in the matching test setup to the finial amplifier bias conditions the better the resulting match can be.
For investigating peak power and current limits in power tubes pulse or fast sweep based measurements are the only safe and practical way to collect the data.
For example the data sheet plate curves on a 6BQ5 shows 165mA at at plate voltage of 400V. That is 66 watts plate dissipation!
The poor 6BQ5 will not likely survive any attempt to confirm this data sheet value with a continuous measurement.
It is my understanding that long ago many (most?) old school tube analyzers used sweep plate voltage measurements with a low duty cycle to probe the upper power limits of tubes.
So depending on the task at hand I find both types of measurements to be useful.
This sound similar to the front end of my build.
Would really like to see your schematic to see how we each realized the finial circuits.
Take care
Interesting solution with the tube+transistor cascode in the 1st stage. I used a similar topology, although with all tubes.
I think this is what got built. The two 'screen' voltages( the FQP1N60 gates ) are adjustable to tune the idle currents of the front end.
Douglas
Attachments
I have too. They're called 'pentodes'... LOL I first built with 6AU6's, and those worked quite well. 6CL6, 6V6, 12BY7, and few more. It seems the variable g2 current matters a bit less than I suspected. EF184 is also good. Wish I had a stash of 7788 to play with.
Douglas
I understand the BJT base current reduces the gain a bit but by such a small amount I feel it can be ignored.
The base emitter resistance also reduces the gain as it allows a tiny signal on the lower triode that reduces gain however the amount of loss with a BJT is very small.
Oh I think I see my error
The total gain is largely a function of S in mA/V of the lower device multiplied by the top load impedance.
Math was never my strong point.
Attachments
Last edited:
This is a interesting circuit you have.
I will have to spice it to see fully how it works.
Thanks for sharing.
Enjoy spicing. I would love to see the results, especially if you could try a few other MOSFET's. I did a fairy exhaustive search when the FQP1N60 was current, and a smaller one when this amp came into being and did not find any I liked better. The CCS tail load is an IXTP08N50D2 over a DN3545N3 cascode.
Douglas
Douglas
The still current STP3LN80K5 maybe worth a look as a replacement for the now hard to find FQP1N60.
The input capacitance is even lower at 102pF and the transfer capacitance is listed at only 0.1pF.
I have a few samples of the STP3LN80K5 but not yet done a lot of testing.
I keep meaning to make up a test jig to evaluate these devices but it keeps getting put off.
The FQP1N60 may have higher capacitances, but by 10V d-s, they are basically done changing. The STP3LN80K5 has a lot of variation above 10V d-s.
Douglas