Hi everyone,
After reading Broskie's hybrid CF+ post here, I decided to play around with the 4P1L triode-strapped DHT and LTspice to see how this circuit may perform for my Grado SR80 headphones which are 32 ohm.
In the attached simulation you can see that simply with one tube running at about 20mA you could get around 0.21% THD when output level is 223Vrms at the 32ohm load. With a sensitivity of 89dB/mW, this would be more than enough as maximum level or am I doing the maths the wrong way here?
This will need just 770Vrms at the input, so will need to add a stepped attenuator and will get a simple headphone amp with this setup.
Any suggestions before attempting a breadboard of this circuit?
Thanks
Ale
After reading Broskie's hybrid CF+ post here, I decided to play around with the 4P1L triode-strapped DHT and LTspice to see how this circuit may perform for my Grado SR80 headphones which are 32 ohm.
In the attached simulation you can see that simply with one tube running at about 20mA you could get around 0.21% THD when output level is 223Vrms at the 32ohm load. With a sensitivity of 89dB/mW, this would be more than enough as maximum level or am I doing the maths the wrong way here?
This will need just 770Vrms at the input, so will need to add a stepped attenuator and will get a simple headphone amp with this setup.
Any suggestions before attempting a breadboard of this circuit?
Thanks
Ale
Attachments
Hi Artosalo,
You can get them from Dmitry's website here:
http://www.dmitrynizh.com/composite-triode-models.htm
Have used the 4P1L with excellent results though, see my thread on 4P1L. Also tweaked your shunt regulator that I have on my 26 preamp to fit the current needs of the 4P1L and worked incredibly well. I'm now doing a side by side comparison!
Any suggestions on this preamp?
Cheers,
Ale
You can get them from Dmitry's website here:
http://www.dmitrynizh.com/composite-triode-models.htm
Have used the 4P1L with excellent results though, see my thread on 4P1L. Also tweaked your shunt regulator that I have on my 26 preamp to fit the current needs of the 4P1L and worked incredibly well. I'm now doing a side by side comparison!
Any suggestions on this preamp?
Cheers,
Ale
4 pole DHT spice symbol
Dmitry's (composite) library - http://www.dmitrynizh.com/dmitry_composites.zip - not compatible with LTSpice five pole triode-with heater symbols.
Use this four pole symbol:
Version 4
SymbolType CELL
LINE Normal -48 0 -28 0
LINE Normal -20 0 -12 0
LINE Normal -4 0 4 0
LINE Normal 12 0 20 0
LINE Normal 28 0 36 0
LINE Normal 0 -48 0 -16
LINE Normal -20 -16 20 -16
LINE Normal -20 -12 20 -12
LINE Normal -20 -16 -20 -12
LINE Normal 20 -16 20 -12
LINE Normal 16 32 16 65
LINE Normal -16 33 -16 65
LINE Normal 1 24 16 32
LINE Normal -16 33 1 24
CIRCLE Normal -48 -48 48 48
WINDOW 0 8 -64 Left 0
WINDOW 3 62 -29 Left 0
SYMATTR Value TriodeDHT
SYMATTR Prefix X
SYMATTR Description This symbol is for use with a subcircuit macromodel that you supply.
PIN 0 -48 NONE 0
PINATTR PinName P
PINATTR SpiceOrder 1
PIN -48 0 NONE 0
PINATTR PinName G
PINATTR SpiceOrder 2
PIN -16 64 NONE 0
PINATTR PinName K1
PINATTR SpiceOrder 3
PIN 16 64 NONE 8
PINATTR PinName K2
PINATTR SpiceOrder 4
Save it to ...lib/sym/misc directory (my symbol name is TriodeDHT.sym).
Dmitry's (composite) library - http://www.dmitrynizh.com/dmitry_composites.zip - not compatible with LTSpice five pole triode-with heater symbols.
Use this four pole symbol:
Version 4
SymbolType CELL
LINE Normal -48 0 -28 0
LINE Normal -20 0 -12 0
LINE Normal -4 0 4 0
LINE Normal 12 0 20 0
LINE Normal 28 0 36 0
LINE Normal 0 -48 0 -16
LINE Normal -20 -16 20 -16
LINE Normal -20 -12 20 -12
LINE Normal -20 -16 -20 -12
LINE Normal 20 -16 20 -12
LINE Normal 16 32 16 65
LINE Normal -16 33 -16 65
LINE Normal 1 24 16 32
LINE Normal -16 33 1 24
CIRCLE Normal -48 -48 48 48
WINDOW 0 8 -64 Left 0
WINDOW 3 62 -29 Left 0
SYMATTR Value TriodeDHT
SYMATTR Prefix X
SYMATTR Description This symbol is for use with a subcircuit macromodel that you supply.
PIN 0 -48 NONE 0
PINATTR PinName P
PINATTR SpiceOrder 1
PIN -48 0 NONE 0
PINATTR PinName G
PINATTR SpiceOrder 2
PIN -16 64 NONE 0
PINATTR PinName K1
PINATTR SpiceOrder 3
PIN 16 64 NONE 8
PINATTR PinName K2
PINATTR SpiceOrder 4
Save it to ...lib/sym/misc directory (my symbol name is TriodeDHT.sym).
Here's a simple model of 4p1l, made with curvecaptor from curves of a 4p1l triode connected. This model does not include capacitance and does not use the cathode as heater. But in ltspice it's easy to use just as any triode. Works fine for playing around with various operating points, distortion, etc.
.subckt 4p1l P G K
Bp P K I=(0.3189882589m)*uramp(V(P,K)*ln(1.0+exp((12.89638275)+(12.89638275)*(10.11832999)*V(G,K)/V(P,K)))/(12.89638275))**(1.120990401)
.ends 4p1l
.subckt 4p1l P G K
Bp P K I=(0.3189882589m)*uramp(V(P,K)*ln(1.0+exp((12.89638275)+(12.89638275)*(10.11832999)*V(G,K)/V(P,K)))/(12.89638275))**(1.120990401)
.ends 4p1l
Today I finished the breadboard version of this amplifier. I decided to parallel two 4P1L to achieve lower output impedance and minimise distortion. In LTspice it performs really well: THD=0.016% @ Vo=360mVrms.
The bipolar capacitors are OSCONs and the value was chosen due to the availability (the -3dB corner is around 5Hz).
Haven't used in this breadboard the Rod Coleman's regulators. Just a good regulated DC supply.
My old bench HT power supply has a significant ripple when dragging 40mA so ended up adding a 470uF electrolytic to minimise the 100Hz bump. See attached.
When testing this circuit, I lowered the HT to 90V to achieve 40.8mA of idle current across the two triodes. When driving it with a 1Vrms I got THD=0.026% @ 484mVrms (1.52Vpp). When tested with the headphones, maximum level for me was around 80mvrms, so distortion is really good.
No ringing issues in CF mode. A bare minimum when you tap the valves. Remember that my grados are 89dB/mW
Sound is warm. If you run the amp at lower than 40mA you get the third harmonic higher than the second. To me the PNP transistor doesn't seem to add any colouration to the sound. It's pleasant to listen to and quiet. This can be improved with a proper HT supply to minimise the ripple
I need to build this properly now...
The bipolar capacitors are OSCONs and the value was chosen due to the availability (the -3dB corner is around 5Hz).
Haven't used in this breadboard the Rod Coleman's regulators. Just a good regulated DC supply.
My old bench HT power supply has a significant ripple when dragging 40mA so ended up adding a 470uF electrolytic to minimise the 100Hz bump. See attached.
When testing this circuit, I lowered the HT to 90V to achieve 40.8mA of idle current across the two triodes. When driving it with a 1Vrms I got THD=0.026% @ 484mVrms (1.52Vpp). When tested with the headphones, maximum level for me was around 80mvrms, so distortion is really good.
No ringing issues in CF mode. A bare minimum when you tap the valves. Remember that my grados are 89dB/mW
Sound is warm. If you run the amp at lower than 40mA you get the third harmonic higher than the second. To me the PNP transistor doesn't seem to add any colouration to the sound. It's pleasant to listen to and quiet. This can be improved with a proper HT supply to minimise the ripple
I need to build this properly now...
Attachments
Last edited:
Wavebourn, hello! do you happen to have a working pentode model of 4P1L for LTspice? Recently I purchased some Chinese 4p1l and want to try them in Schaded connection.
Wavebourn, hello! do you happen to have a working pentode model of 4P1L for LTspice? Recently I purchased some Chinese 4p1l and want to try them in Schaded connection.
Yes, I have it in my head, but it is incompatible with SPICE. Works fine, though.
It should work in pentode with local feedback like a charm. But G2 is too dence, it loks like it was made from a nice triodeprototype when a boss ordered developer - audiophile to make a pentode for military.
Yes, I have it in my head, but it is incompatible with SPICE. Works fine, though.
It should work in pentode with local feedback like a charm. But G2 is too dence, it loks like it was made from a nice triodeprototype when a boss ordered developer - audiophile to make a pentode for military.
No problem, I will give it a shot
Hi Wavebourn, some quick questions, what is the voltage swing seeing by the 4p1l grid to have 2.5W at output?
Are we running into class A2? what is approximate idle current?
Hi Wavebourn, some quick questions, what is the voltage swing seeing by the 4p1l grid to have 2.5W at output?
In triode mode? Somerhing like -18V bias on G1, and up to +12V swing. Either transformer, or cathode/source follower is needed. About 50 mA, but less than causes 11W dissipation (anode + G2).
However, in pentode mode things will be different. Your call.
Hint: after WW-II during "Battle for harwest", some equipment was made for agriculture using military devices, and a collective-farm radio receiver had 4P1L in driver, and 2 of them in P-P output, with interstage transformer. Output power was 10W. Plenty for the willage.
The concept here I think is a good one, an all DHT headamp. But probably a tad underpowered. But maybe a better (albeit more expensive) means to this end would be a typical 45 SET type design. Only for the gain stage we use another 4p1l. Originally I wanted a #26 or #112, but why not use a triode 4p1l for the gain and one for the driver. Total gain would be 8x8=64 (as much as the defacto D3A/6p45c spud SETs we are all using). If it could output a clean 2Vrms to a pair of Fostex TP50RP's it might be worth trying. Heck it might even put 1W into an 8 ohm tap. I wonder how low the 4P1L Rp could go with the right operating point, certainly better than a 45 even the $300 ones. What do folks think, is this worth pursuing ?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.