Cathode Driven Tube Output Stages
.
Just in case anyone following the thread is unaware, Cathode Driven P-P Tube Output stages have been well documented and described. Not new.
Some cat named Wiggins seems to have done quite a few things in that area as well... 😉
The erstwhile and venerable Radiotron Designer's Handbook is a good reference to explore that domain.
😀
_-_-bear
PS. interstage driver transformers that drive the grids of tubes tend to be step down, not step up so that they can deliver the requisite power needed to get into A2 or AB2... an interesting detail of difference between the soylent state and the glow fet state device's operation.
.
Just in case anyone following the thread is unaware, Cathode Driven P-P Tube Output stages have been well documented and described. Not new.
Some cat named Wiggins seems to have done quite a few things in that area as well... 😉
The erstwhile and venerable Radiotron Designer's Handbook is a good reference to explore that domain.
😀
_-_-bear
PS. interstage driver transformers that drive the grids of tubes tend to be step down, not step up so that they can deliver the requisite power needed to get into A2 or AB2... an interesting detail of difference between the soylent state and the glow fet state device's operation.
Re: Cathode Driven Tube Output Stages
Dear Bear,
Thank you for your timely contribution, pertinent as always 🙂
I was asked about a tube version of zeus - this is what it could look like.
I am sure we would all be very appreciative if you could be so kind as to quote chapter and page number for these references.
Zeus operates in Class A.
Step down is used to match impedances between tubes that are amplification stages.
Step up interstage transformer where the amplification is in the inductor and the tubes are followers.
The tube version, despite it's similar appearance to the mosfet circuit, is not simple to implement and has much higher voltages and is therefor not suitable as a DIY beginner's project.
Best wishes,
Susan.
Dear Bear,
Thank you for your timely contribution, pertinent as always 🙂
bear said:
I was asked about a tube version of zeus - this is what it could look like.
I am sure we would all be very appreciative if you could be so kind as to quote chapter and page number for these references.
Zeus operates in Class A.
Step down is used to match impedances between tubes that are amplification stages.
Step up interstage transformer where the amplification is in the inductor and the tubes are followers.
The tube version, despite it's similar appearance to the mosfet circuit, is not simple to implement and has much higher voltages and is therefor not suitable as a DIY beginner's project.
Best wishes,
Susan.
Air Core Output Transformer
Dear All,
I hope everyone is keeping well and had a good holiday.
I just happend to have a reel of cable that I though might be suitable (as a starting point) for an air core transformer so I decided to "give it a go".
This is what it looks like 🙂
100 meters of twin axial cable...
Specifications:
==========
100 metres of twin axial 100 ohm cable, RS part number: 388-316
Inductance: 3.85 mH @ 1 kHz (4.82 mH at 120 Hz).
Capacitance: 4.9 nF conductor to conductor, 7 nF conductor to screen.
Impedance of one winding: 80 ohms at 1 kHz.
Nominal Cable Impedance: 100 ohm.
Capacitance per meter: 50.9 pF
Cores: 7/0.3 mm insulated with polythene, + overall polythene sheath.
Braid: Copper, tinned.
Weight: 11 kg
Price: UKP 92 + VAT
Wiring: 7/03 conductors connected as two arms of the primaries. The braid as secondary to speaker. I.e. 2:1 step down.
2SK1529 Lateral FETs
VBias 5.0 volts, for 1.6 amps quiescent (both fets).
THD @ 1kHz: 8 watts = 0.120%, 2 watts = 0.033%, 1 watt = 0.020%
I have only done a few tests (I will post the results seperatly) and had a quick listen but "Dire Straights" sounds okay. Difficult to say if better than with a conventional transformer but it is a promising start.
Needs more testing and refining but it does work (somewhat to my surprise) with practical sized (i.e. something that I can actually lift) coil.
Best wishes,
Susan.
Dear All,
I hope everyone is keeping well and had a good holiday.
I just happend to have a reel of cable that I though might be suitable (as a starting point) for an air core transformer so I decided to "give it a go".
This is what it looks like 🙂
100 meters of twin axial cable...
Specifications:
==========
100 metres of twin axial 100 ohm cable, RS part number: 388-316
Inductance: 3.85 mH @ 1 kHz (4.82 mH at 120 Hz).
Capacitance: 4.9 nF conductor to conductor, 7 nF conductor to screen.
Impedance of one winding: 80 ohms at 1 kHz.
Nominal Cable Impedance: 100 ohm.
Capacitance per meter: 50.9 pF
Cores: 7/0.3 mm insulated with polythene, + overall polythene sheath.
Braid: Copper, tinned.
Weight: 11 kg
Price: UKP 92 + VAT
Wiring: 7/03 conductors connected as two arms of the primaries. The braid as secondary to speaker. I.e. 2:1 step down.
2SK1529 Lateral FETs
VBias 5.0 volts, for 1.6 amps quiescent (both fets).
THD @ 1kHz: 8 watts = 0.120%, 2 watts = 0.033%, 1 watt = 0.020%
I have only done a few tests (I will post the results seperatly) and had a quick listen but "Dire Straights" sounds okay. Difficult to say if better than with a conventional transformer but it is a promising start.
Needs more testing and refining but it does work (somewhat to my surprise) with practical sized (i.e. something that I can actually lift) coil.
Best wishes,
Susan.
Air Core Output Transformer II
Dear All,
Here are the figures for the HP 8903B tests I did yesterday, THD% for 8 and 2 watts.
Hz 8W-THD 2W-THD
100 4.74 0.283
150 2.533 0.117
200 1.462 0.076
250 0.826 0.058
300 0.454 0.050
400 0.209 0.042
500 0.157 0.039
600 0.139 0.036
700 0.131 0.034
800 0.126 0.034
900 0.123 0.033
1000 0.120 0.033
1200 0.115 0.032
1500 0.111 0.030
2000 0.105 0.029
2500 0.100 0.028
3000 0.097 0.028
4000 0.089 0.027
5000 0.082 0.027
6000 0.074 0.026
7000 0.065 0.025
8000 0.055 0.025
9000 0.046 0.024
10000 0.038 0.024
11000 0.032 0.024
12000 0.032 0.025
15000 0.066 0.029
20000 0.153 0.051
25000 0.251 0.084
30000 0.360 0.110
40000 0.782 0.225
50000 1.138 0.391
60000 1.077 0.424
70000 1.026 0.360
80000 1.690 0.486
90000 2.003 0.742
100000 2.059 0.834 -3dB point.
2SK1529 Lateral FETs VBias at 5.0 volts, for 1.6 amps quiescent (both fets).
This is for the circuit as shown in the picture with four transformers and two power stages.
Best wishes,
Susan.
Dear All,
Here are the figures for the HP 8903B tests I did yesterday, THD% for 8 and 2 watts.
Hz 8W-THD 2W-THD
100 4.74 0.283
150 2.533 0.117
200 1.462 0.076
250 0.826 0.058
300 0.454 0.050
400 0.209 0.042
500 0.157 0.039
600 0.139 0.036
700 0.131 0.034
800 0.126 0.034
900 0.123 0.033
1000 0.120 0.033
1200 0.115 0.032
1500 0.111 0.030
2000 0.105 0.029
2500 0.100 0.028
3000 0.097 0.028
4000 0.089 0.027
5000 0.082 0.027
6000 0.074 0.026
7000 0.065 0.025
8000 0.055 0.025
9000 0.046 0.024
10000 0.038 0.024
11000 0.032 0.024
12000 0.032 0.025
15000 0.066 0.029
20000 0.153 0.051
25000 0.251 0.084
30000 0.360 0.110
40000 0.782 0.225
50000 1.138 0.391
60000 1.077 0.424
70000 1.026 0.360
80000 1.690 0.486
90000 2.003 0.742
100000 2.059 0.834 -3dB point.
2SK1529 Lateral FETs VBias at 5.0 volts, for 1.6 amps quiescent (both fets).
This is for the circuit as shown in the picture with four transformers and two power stages.
Best wishes,
Susan.
Hi Susan,
Does that reel of screened twin have a 'shorting' centre turn of steel ?
If yes, maybe it can be removed for better low frequency response.
Put a large old transformer in the hole ?
Ever thought of having two Mosfets in parallel per half - one low transconductance, the other high.
Say add high transconductance but low voltage bias devices to your existing setup using low value source resistors. This might maintain a minimum conduction of class-A, say 100mA operation when the main transistors are approaching cut-off. Hope you can understand my drift. I am thinking about a gradual overlap of different characteristics, but it might turn out to be superfluous.
Cheers ......... Graham.
Does that reel of screened twin have a 'shorting' centre turn of steel ?
If yes, maybe it can be removed for better low frequency response.
Put a large old transformer in the hole ?
Ever thought of having two Mosfets in parallel per half - one low transconductance, the other high.
Say add high transconductance but low voltage bias devices to your existing setup using low value source resistors. This might maintain a minimum conduction of class-A, say 100mA operation when the main transistors are approaching cut-off. Hope you can understand my drift. I am thinking about a gradual overlap of different characteristics, but it might turn out to be superfluous.
Cheers ......... Graham.
Twinax Cable
Dear Susan,
Happy New Year !!
a) I am glad that the 2SK's are still going strong.
b) That twinax stuff you have is great as balanced interconnects (XLR). I have been trying to get some but could not convince myself to buy 100m when I only need 3. So when you have eventually finished with the test, you may well consider offering them on the internet as inteconnects by the metre, or at least give me a shout before you chuck them away !! 🙂
Patrick
Dear Susan,
Happy New Year !!
a) I am glad that the 2SK's are still going strong.
b) That twinax stuff you have is great as balanced interconnects (XLR). I have been trying to get some but could not convince myself to buy 100m when I only need 3. So when you have eventually finished with the test, you may well consider offering them on the internet as inteconnects by the metre, or at least give me a shout before you chuck them away !! 🙂
Patrick
Hi Graham,
Thanks for your post 🙂
There is a steel center, but it is only about 300 degrees i.e. there is a two inch slot.
I tried the cable on an off chance and was not really expecting it to work down much below 10 kHz.
However the inductance at 1 kHz is 3.85 mH which is higher than I was expecting.
Next is to remove that bit of steel and see how much is affects the inductance.
I was thinking of a speaker cone... 🙂
I have looked at various configurations but they all add complexity...
... but certainly something one can experiment with.
Best wishes,
Susan.
Thanks for your post 🙂
Graham Maynard said:
There is a steel center, but it is only about 300 degrees i.e. there is a two inch slot.
I tried the cable on an off chance and was not really expecting it to work down much below 10 kHz.
However the inductance at 1 kHz is 3.85 mH which is higher than I was expecting.
Next is to remove that bit of steel and see how much is affects the inductance.
I was thinking of a speaker cone... 🙂
I have looked at various configurations but they all add complexity...
... but certainly something one can experiment with.
Best wishes,
Susan.
Re: Twinax Cable
Hi Patrick,
Thanks for your post 🙂
Best wishes for 2005.
Yes, and the 2SK's are running up higher with the lower inductance.
Much as I would like to be able to help the coil of twinax has now become an official zeus transformer.
The question is whether to buy another 100 meter reel and make up a 4:1 transformer.
Guess I will have to do that to be able to also set up a stereo pair for proper listening tests.
Best wishes,
Susan.
Hi Patrick,
Thanks for your post 🙂
Best wishes for 2005.
EUVL said:
Yes, and the 2SK's are running up higher with the lower inductance.
Much as I would like to be able to help the coil of twinax has now become an official zeus transformer.
The question is whether to buy another 100 meter reel and make up a 4:1 transformer.
Guess I will have to do that to be able to also set up a stereo pair for proper listening tests.
Best wishes,
Susan.
Simple components question...
Hi Susan,
I am ordering a bunch of parts for other compnents and was wondering what the voltage rating should be on the capacitors(10,000uf, 100n/0.1uf, 4.7uf)) in the high current version of the Zeus?
I'm currently looking at 10,000uf/63V Nichicon KG Gold($14 US) for the main power supply and for the bias Black Gate NX Hi-Q 0.1uf/50V ($1.35) or Rel RTX 0.1uf/400V ($6.95), and Black gate N 4.7uf/50V.
Do I need to get 100V caps or higher for safety?
I will use IRC 1/4W 0.1% precision resistors.
Any new info on the prewound input and bare output bobbins?
I'll check with Thomas & Skinner on the laminations costs.
Hi Susan,
I am ordering a bunch of parts for other compnents and was wondering what the voltage rating should be on the capacitors(10,000uf, 100n/0.1uf, 4.7uf)) in the high current version of the Zeus?
I'm currently looking at 10,000uf/63V Nichicon KG Gold($14 US) for the main power supply and for the bias Black Gate NX Hi-Q 0.1uf/50V ($1.35) or Rel RTX 0.1uf/400V ($6.95), and Black gate N 4.7uf/50V.
Do I need to get 100V caps or higher for safety?
I will use IRC 1/4W 0.1% precision resistors.
Any new info on the prewound input and bare output bobbins?
I'll check with Thomas & Skinner on the laminations costs.
Re: Simple components question...
Hi Darkmoebius,
🙂
I would use a good quality electrolytic plus a metallised polyester across it for the bias as well.
What is your proposed power rail voltage?
I am planning on having about a 40 volt supply rail so will be using 63 volt caps.
One thing one has to be careful of is nominal line voltage versus real. Here in the UK we have a 230 Vac normalized to the rest of Europe, so all the transformers are now 230 volts instead of 240 volts.
Great, except our local supply in London is 245 Vac, which rather messes with the calculations.
Plus spikes etc. I also use those transient spike suppression things on the input (after the fuse) across the primaries. As I use 120 + 120 windings so the amp can be moved around (my originals went for a couple of world tours) these also will blow the fuse if the amp is in 120 mode and inadvertently connected to a 240 volt supply.
63 volt rated should be okay if you are running with under 40 volts on the DC rail but if it was getting close to 50 volts I would probably go for the higher voltage rating to allow for surges etc.
Bias should never be over 5 or 6 volts
Should be fine. No point in going much beyond that though as the mosfet matching won't be exact.
Saying that I have added a pack of 5 to an RS order and I will compare those to the ones I am using and see if there is any noticable difference.
I am sorting this out, however I am just a little distracted at the moment by xml keyboard layout definition files for OS X (running on an iBook that has the tendency to crash every time I open the system preferences - I have had to reinstall the system twice so far).
Will send some emails 🙂
Best wishes,
Susan.
Hi Darkmoebius,
🙂
darkmoebius said:
I would use a good quality electrolytic plus a metallised polyester across it for the bias as well.
What is your proposed power rail voltage?
I am planning on having about a 40 volt supply rail so will be using 63 volt caps.
One thing one has to be careful of is nominal line voltage versus real. Here in the UK we have a 230 Vac normalized to the rest of Europe, so all the transformers are now 230 volts instead of 240 volts.
Great, except our local supply in London is 245 Vac, which rather messes with the calculations.
Plus spikes etc. I also use those transient spike suppression things on the input (after the fuse) across the primaries. As I use 120 + 120 windings so the amp can be moved around (my originals went for a couple of world tours) these also will blow the fuse if the amp is in 120 mode and inadvertently connected to a 240 volt supply.
63 volt rated should be okay if you are running with under 40 volts on the DC rail but if it was getting close to 50 volts I would probably go for the higher voltage rating to allow for surges etc.
Bias should never be over 5 or 6 volts
Should be fine. No point in going much beyond that though as the mosfet matching won't be exact.
Saying that I have added a pack of 5 to an RS order and I will compare those to the ones I am using and see if there is any noticable difference.
I am sorting this out, however I am just a little distracted at the moment by xml keyboard layout definition files for OS X (running on an iBook that has the tendency to crash every time I open the system preferences - I have had to reinstall the system twice so far).
Will send some emails 🙂
Best wishes,
Susan.
Re: Simple components question...
Hi Darkmoebius,
Further to previous post...
Did a quick check, and I can confirm that the mosfet gate resistor matching is not super critical.
If one is looking at specs then low inductance and low thermal noise over super close matching.
Best wishes,
Susan.
Hi Darkmoebius,
Further to previous post...
darkmoebius said:
Did a quick check, and I can confirm that the mosfet gate resistor matching is not super critical.
If one is looking at specs then low inductance and low thermal noise over super close matching.
Best wishes,
Susan.
Zeus 75 Prototype
Dear All.
Constructed with a Sowter wound bobbin (to confirm wiring details).
Chassis is 1.5mm steel.
Heatsink is 250mm high.
Note: heatsink is spaced off from the chassis with a washer at each bolt.
Transformer leads left long for testing.
Mosfets are STW34NB20s
Bias using ST L200 in current regulator mode (6R8 series resistor for 100 mA), Schottky diode and third STW34NB20.
Best wishes,
Susan.
Dear All.
Constructed with a Sowter wound bobbin (to confirm wiring details).
Chassis is 1.5mm steel.
Heatsink is 250mm high.
Note: heatsink is spaced off from the chassis with a washer at each bolt.
Transformer leads left long for testing.
Mosfets are STW34NB20s
Bias using ST L200 in current regulator mode (6R8 series resistor for 100 mA), Schottky diode and third STW34NB20.
Best wishes,
Susan.
Very nice!!!
Hi Susan,
Can't get any more clean, compact, and simple than that. It looks much smaller than I pictured in my mind.
There are two XLR connectors - Is one power and the other signal?
And, why two sets of speaker posts?
Hi Susan,
Can't get any more clean, compact, and simple than that. It looks much smaller than I pictured in my mind.
There are two XLR connectors - Is one power and the other signal?
And, why two sets of speaker posts?
Re: Very nice!!!
Answered my own question on closer inspection.
Aaaah, I suppose 8 and 16 ohm taps. Or 4 & 8 ohm.
Man, I really like this design. Bravo Susan!
darkmoebius said:
Answered my own question on closer inspection.
Aaaah, I suppose 8 and 16 ohm taps. Or 4 & 8 ohm.
Man, I really like this design. Bravo Susan!
A tubeless tube amp! 
Re: Very nice!!!
Hi Darkmoebius ,
Thanks.
It's about 12" wide, 9-1/4" deep and 10-1/4" high, and according to my bathroom scales (probably not the most accurate of measures) weighs 29 lbs / 13.5 kgs.
Those handles are NOT decorative 🙂
The amp can sit on it's power supply (as shown in the 3D sketch) and live behind or to one side of the speaker it is driving.
Yes, the 4 pin XLR is for power, but may be replaced by a touchproof power connector (that fits the "universal" XLR chassis cutout) for the DC power interconnect between the PSU and the amp.
( Cliff 4 pole touchproof and locking connectors to IEC65/IEC348.)
For the PSU I am also looking at the Neutrik AG Powercon NAC3MPA (chassis) and NAC3FCA (cable) for line ac power.
There are two pairs of posts so one can configure the output series for 2:1 or parallel for 4:1 step down, or use the two 4:1 sections separately to bi-wire to a speaker (the tweeter and bass/mid then have their own set of windings on the transformer).
"Production" amps will probably have a 4 way Speakon connector. I would like to use the Multi-contact touchproof 6 mm connectors, but they are very expensive.
Thank you 🙂
Now I have to build the power supply.
Best wishes,
Susan.
Hi Darkmoebius ,
darkmoebius said:
Thanks.
It's about 12" wide, 9-1/4" deep and 10-1/4" high, and according to my bathroom scales (probably not the most accurate of measures) weighs 29 lbs / 13.5 kgs.
Those handles are NOT decorative 🙂
The amp can sit on it's power supply (as shown in the 3D sketch) and live behind or to one side of the speaker it is driving.
Yes, the 4 pin XLR is for power, but may be replaced by a touchproof power connector (that fits the "universal" XLR chassis cutout) for the DC power interconnect between the PSU and the amp.
( Cliff 4 pole touchproof and locking connectors to IEC65/IEC348.)
For the PSU I am also looking at the Neutrik AG Powercon NAC3MPA (chassis) and NAC3FCA (cable) for line ac power.
There are two pairs of posts so one can configure the output series for 2:1 or parallel for 4:1 step down, or use the two 4:1 sections separately to bi-wire to a speaker (the tweeter and bass/mid then have their own set of windings on the transformer).
"Production" amps will probably have a 4 way Speakon connector. I would like to use the Multi-contact touchproof 6 mm connectors, but they are very expensive.
Thank you 🙂
Now I have to build the power supply.
Best wishes,
Susan.
Hi Hugo,
Thanks...
... Acoustic guitar sounds very nice 🙂
Generally same sort of sound as I have been getting before but I have only had a quick listen as I got the amp running yesterday evening (I then had to go and cook supper).
These mosfets are matched to within a couple of mV at 750 mA bias each (I got 3 sets of good matches from 10 parts).
4:1 step down, TPA line driver, 4.6 watts into 8 ohms, 1kHz peak at -2.6 dB on scale, 0.0083% THD.
Just a quick check, not set up properly.
Note the reduction in the mains line power 50 Hz and harmonics.
Best wishes,
Susan.
Netlist said:
Thanks...
... Acoustic guitar sounds very nice 🙂
Generally same sort of sound as I have been getting before but I have only had a quick listen as I got the amp running yesterday evening (I then had to go and cook supper).
These mosfets are matched to within a couple of mV at 750 mA bias each (I got 3 sets of good matches from 10 parts).
An externally hosted image should be here but it was not working when we last tested it.
4:1 step down, TPA line driver, 4.6 watts into 8 ohms, 1kHz peak at -2.6 dB on scale, 0.0083% THD.
Just a quick check, not set up properly.
Note the reduction in the mains line power 50 Hz and harmonics.
Best wishes,
Susan.