Susan, from your website:
I have experienced the same effect, however when I short the power supply cap so that there is absolutely no charge left in it the amplifier goes quiet.
I believe the amplifier makes the distorted sound because the bias voltage drops, thus creating heavy cross-over distortion.
I have experienced the same effect, however when I short the power supply cap so that there is absolutely no charge left in it the amplifier goes quiet.
I believe the amplifier makes the distorted sound because the bias voltage drops, thus creating heavy cross-over distortion.
How about using Susan's topology for portable headphone amplification?
I'd imagine the 2SK389 transistors would cut down on the work required for matching, while also providing the power to drive a set of not-too-current-hungry headphones.
Alternately, if anyone has a good source for matched pairs of MGF1302 (GaAsFET's, about USD 10 ea, so matching get's expensive rather quickly) they can deliver up to 100mA each from a single D cell.
The transformers for something like this should also be relatively inexpensive.
I'd imagine the 2SK389 transistors would cut down on the work required for matching, while also providing the power to drive a set of not-too-current-hungry headphones.
Alternately, if anyone has a good source for matched pairs of MGF1302 (GaAsFET's, about USD 10 ea, so matching get's expensive rather quickly) they can deliver up to 100mA each from a single D cell.
The transformers for something like this should also be relatively inexpensive.
Hi Bobo1on1,
Thank you for your post 🙂
And when you remove the short the amplifier will once more make a noise.
There is no bias as such, simply the charge generated from the input signal which produces enough to cause the mosfet to conduct.
I have 10K resistors across my power supply capacitors so the volts dissipate quite quickly.
When an input signal is applied (and it has to be of reasonable level) the main power rail gains a hundred mV or so, but the gate goes up several volts (to the mosfet gate offset voltage).
This pic shows the effect with the SE amplifier.
Best wishes,
Susan.
P.S.
Some further updates to the SE page
http://www.susan-parker.co.uk/zeus-se-amp.htm
and yes I have been listening to it as well as taking measurements.
Thank you for your post 🙂
bobo1on1 said:
And when you remove the short the amplifier will once more make a noise.
There is no bias as such, simply the charge generated from the input signal which produces enough to cause the mosfet to conduct.
I have 10K resistors across my power supply capacitors so the volts dissipate quite quickly.
When an input signal is applied (and it has to be of reasonable level) the main power rail gains a hundred mV or so, but the gate goes up several volts (to the mosfet gate offset voltage).
This pic shows the effect with the SE amplifier.
Best wishes,
Susan.
P.S.
Some further updates to the SE page
http://www.susan-parker.co.uk/zeus-se-amp.htm
and yes I have been listening to it as well as taking measurements.
The problem of NFB ampli, is that you' ll have a courrent source like ampli that will sound ok only if the speaker connected have a linear resistance in all the spectrum.
I also tried to play with NFB.....very good results with magneplanar isodinamic louspeaker http://www.magnepan.com/index.php because their impedance module is quite a perfect line in all the audible spectrum.
Another good try is to use some Alnico or ferrofluid louspeaker that have a better impedance module linearity, but really don't tried it cause they are veeeeery expensives.
If you' ll try NFB you'll love it (with the right loudspeaker)
I also tried to play with NFB.....very good results with magneplanar isodinamic louspeaker http://www.magnepan.com/index.php because their impedance module is quite a perfect line in all the audible spectrum.
Another good try is to use some Alnico or ferrofluid louspeaker that have a better impedance module linearity, but really don't tried it cause they are veeeeery expensives.
If you' ll try NFB you'll love it (with the right loudspeaker)
Hi Rinox,
Thank you for your post 🙂
With the 4:1 step down configuration the output impedance is only a quarter of an ohm, which gives a damping factor of over 30.
Yes, there is some variation in amplitude relative to frequency with ordinary speakers but it is small enough in comparison to room effects not to be (for me anyway) an issue.
As you say, a speaker with a flat impedance curve would seem to be ideal. The magneplanars are typically 86 dB at 2.83 Vac, but being 4 ohms this is 2 watts so the push-pull version of the amp would probably be required (the reviews mention needing a powerful amp).
My own experience with the various speakers that I have tried is that they all sound better when driven with my amplifier. Choosing one's that are known to work well with valve amplification is a good start, but not essential.
And matching speakers to amps seems to be the accepted thing even with solid state feedback amplifiers, so I don't see this as a particular problem.
Thoughts?
Best wishes,
Susan.
Thank you for your post 🙂
rinox said:
With the 4:1 step down configuration the output impedance is only a quarter of an ohm, which gives a damping factor of over 30.
Yes, there is some variation in amplitude relative to frequency with ordinary speakers but it is small enough in comparison to room effects not to be (for me anyway) an issue.
As you say, a speaker with a flat impedance curve would seem to be ideal. The magneplanars are typically 86 dB at 2.83 Vac, but being 4 ohms this is 2 watts so the push-pull version of the amp would probably be required (the reviews mention needing a powerful amp).
My own experience with the various speakers that I have tried is that they all sound better when driven with my amplifier. Choosing one's that are known to work well with valve amplification is a good start, but not essential.
And matching speakers to amps seems to be the accepted thing even with solid state feedback amplifiers, so I don't see this as a particular problem.
Thoughts?
Best wishes,
Susan.
Hi Rinox,
But Susan is using NFB !
Its local - not global.
The voltage gain comes from the input transformer;
the current drive from the Mosfets with self sensing voltage feedback that provides a medium level of loudspeaker damping.
(More than I originally expected from just looking at the circuit diagram.)
Unlike other NFB amplifiers her circuit has no reactive components other than the Mosfets themselves within that NFB loop; and here, with a good input transformer plus low impedance pre-amplifier drive, Mosfet reactivity is not significant within the audio spectrum.
Regarding loudspeaker impedance;- complex dynamic loudspeaker systems that have been compensated to give an optimally linear 'resistance' (passive sinewave impedance) throughout the audio spectrum, will still not necessarily reproduce the same with either a predominantly current or voltage amplifier output. This is because the dynamic characteristics of all those components used to produce the flat sinusoidally measured characteristic remain incapable of generating a coherent response to dynamic (music) waveform change and will still (unlike Magneplanars and electrostatics) react quite differently when driven by different amplifier types.
This is where, ( in the absence of Magneplanar/electrostatic loudspeakers) and as already used by Susan, a small good quality full-range driver can reproduce most respectably.
Please do not think I that am biased towards Susan's design; I really am offering honest opinion.
Cheers ............. Graham.
But Susan is using NFB !
Its local - not global.
The voltage gain comes from the input transformer;
the current drive from the Mosfets with self sensing voltage feedback that provides a medium level of loudspeaker damping.
(More than I originally expected from just looking at the circuit diagram.)
Unlike other NFB amplifiers her circuit has no reactive components other than the Mosfets themselves within that NFB loop; and here, with a good input transformer plus low impedance pre-amplifier drive, Mosfet reactivity is not significant within the audio spectrum.
Regarding loudspeaker impedance;- complex dynamic loudspeaker systems that have been compensated to give an optimally linear 'resistance' (passive sinewave impedance) throughout the audio spectrum, will still not necessarily reproduce the same with either a predominantly current or voltage amplifier output. This is because the dynamic characteristics of all those components used to produce the flat sinusoidally measured characteristic remain incapable of generating a coherent response to dynamic (music) waveform change and will still (unlike Magneplanars and electrostatics) react quite differently when driven by different amplifier types.
This is where, ( in the absence of Magneplanar/electrostatic loudspeakers) and as already used by Susan, a small good quality full-range driver can reproduce most respectably.
Please do not think I that am biased towards Susan's design; I really am offering honest opinion.
Cheers ............. Graham.
Who is
Who is Susan Parker ? and why have this person beefed up a transistor radio circuit, I will remember that this topology was used in almost any of the Solid State transistor radios in the dim and distant past, at least almost any radio from europe is this construction a step forward or is it just an experiment.
Kamskoma
Who is Susan Parker ? and why have this person beefed up a transistor radio circuit, I will remember that this topology was used in almost any of the Solid State transistor radios in the dim and distant past, at least almost any radio from europe is this construction a step forward or is it just an experiment.
Kamskoma
kamskoma
If you look carefully you will see that it is different to the old transistor radio circuits in a subtle but important way. The concerns you have were discussed at the beginning of the thread.🙂
If you look carefully you will see that it is different to the old transistor radio circuits in a subtle but important way. The concerns you have were discussed at the beginning of the thread.🙂
Hi Rinox !
Regarding output impedance: If You have ever looked at the frequens-response of a loudspeaker You will know that it is not flat. And the variations caused by a higher drive impedance is not nessesary wrong. It might be better, it might be worse, it all depends.
The only problems with a low damping factor is less damping of, the basic resonanse of the speaker-unit and resonanses from the reactive components of the crossover.
The first "problem" can be used to ones advantage by using a low Q driver in a big closed box and then using the output impedance to get the Q up to a suitable level ( 0,5 to 1 ).
The second problem can be solved by redesign of the crossover. Nelson Pass has written an article about that. You can find it at his "First Watt" homepage. ( I think its called "Current source Amplifiers")
So nothing is right or wrong with current or voltage drive. Current drive will actually reduce some kinds of distortion.
😉 Thorsten
Regarding output impedance: If You have ever looked at the frequens-response of a loudspeaker You will know that it is not flat. And the variations caused by a higher drive impedance is not nessesary wrong. It might be better, it might be worse, it all depends.
The only problems with a low damping factor is less damping of, the basic resonanse of the speaker-unit and resonanses from the reactive components of the crossover.
The first "problem" can be used to ones advantage by using a low Q driver in a big closed box and then using the output impedance to get the Q up to a suitable level ( 0,5 to 1 ).
The second problem can be solved by redesign of the crossover. Nelson Pass has written an article about that. You can find it at his "First Watt" homepage. ( I think its called "Current source Amplifiers")
So nothing is right or wrong with current or voltage drive. Current drive will actually reduce some kinds of distortion.
😉 Thorsten
Hi Kamskoma !
The progress in semiconductor technology has made it possible to return to simpler topologies, but with more power and less distortion than in the old days.
You might like to try a super simple amplifier Yourself. 😉
I was shocked by the nearness of the performance the first time I heard the Zen-amplifier. (Nelsons)
Thorsten
The progress in semiconductor technology has made it possible to return to simpler topologies, but with more power and less distortion than in the old days.
You might like to try a super simple amplifier Yourself. 😉
I was shocked by the nearness of the performance the first time I heard the Zen-amplifier. (Nelsons)
Thorsten
Zeus Portable Headphone Amp
Hi Angel,
Thank you for your post and interesting proposition for headphones 🙂
Sorry for the delay in getting back on this as I was out all day yesterday (keyboards then at IBM Hursley Park in the evening) and wanted to do a little benchwork before replying.
I don't have any 2SK389 so I used some STD5N20s and a pair of little 1:1CT transformers from RS, # 210-6368 - E187A, 800Ω imp, 2mW, 200-15kHz, dc 69Ω/97Ω .
Bias set to 10 mA (for the pair i.e. about 5 mA each - mosfets were not matched).
Supply 12 volts (worked okay down to 10 volts with these mosfets).
With a 600 ohm load at 500 mV ac I got the following:
-3 dB response = 63 Hz to 50 kHz (output at 20% at 100 kHz).
Frequency = THD%
=============
50 Hz = 2.7% - with 80 kHz low pass filter.
63 Hz = 1.9% - with 80 kHz filter.
100 Hz = 0.96% - with 80 kHz filter.
1 kHz = 0.05% - with 80 kHz filter, 0.09% without.
10 kHz = 0.07% - no filter
20 kHz = 0.07% - no filter
30 kHz = 0.17% - no filter
50 kHz = 0.22% - no filter
100 kHz = 0.56% - no filter
N.B. HP 8903B Analyser measurements without filter are with 500 kHz bandwidth input.
The amp would benefit from a slightly larger output transformer (perhaps something with EI 1/2" Mumetal laminations?). However as a quick bench lash up it does sound okay with Sony MDR80 mk IIs.
Current drain is 11.9 mA when driving headphones.
Overall looks promising 🙂
Thanks.
Best wishes,
Susan.
Hi Angel,
Thank you for your post and interesting proposition for headphones 🙂
Sorry for the delay in getting back on this as I was out all day yesterday (keyboards then at IBM Hursley Park in the evening) and wanted to do a little benchwork before replying.
angel said:
I don't have any 2SK389 so I used some STD5N20s and a pair of little 1:1CT transformers from RS, # 210-6368 - E187A, 800Ω imp, 2mW, 200-15kHz, dc 69Ω/97Ω .
Bias set to 10 mA (for the pair i.e. about 5 mA each - mosfets were not matched).
Supply 12 volts (worked okay down to 10 volts with these mosfets).
With a 600 ohm load at 500 mV ac I got the following:
-3 dB response = 63 Hz to 50 kHz (output at 20% at 100 kHz).
Frequency = THD%
=============
50 Hz = 2.7% - with 80 kHz low pass filter.
63 Hz = 1.9% - with 80 kHz filter.
100 Hz = 0.96% - with 80 kHz filter.
1 kHz = 0.05% - with 80 kHz filter, 0.09% without.
10 kHz = 0.07% - no filter
20 kHz = 0.07% - no filter
30 kHz = 0.17% - no filter
50 kHz = 0.22% - no filter
100 kHz = 0.56% - no filter
N.B. HP 8903B Analyser measurements without filter are with 500 kHz bandwidth input.
The amp would benefit from a slightly larger output transformer (perhaps something with EI 1/2" Mumetal laminations?). However as a quick bench lash up it does sound okay with Sony MDR80 mk IIs.
Current drain is 11.9 mA when driving headphones.
Overall looks promising 🙂
Thanks.
Best wishes,
Susan.
Hi,
Thanks 🙂
I changed the transformers to use the Gardners 1:4 for input and an RS 228-393 1+1:2+2 for the output (wired the other way around as a 4:1 step down).
This is the performance into 600 ohms:
Frequency THD% Vac
20.0 1.559 0.444
25.0 0.956 0.455
31.5 0.610 0.464
40.0 0.386 0.472
50.0 0.254 0.478
63.0 0.168 0.484
80.0 0.110 0.488
100 0.077 0.492
125 0.058 0.495
160 0.045 0.497
200 0.038 0.499
250 0.035 0.501
315 0.030 0.503
400 0.027 0.504
500 0.025 0.505
630 0.023 0.506
800 0.021 0.508
1000 0.020 0.508
1250 0.019 0.509
1600 0.019 0.511
2000 0.019 0.511
2500 0.019 0.511
3150 0.019 0.512
4000 0.019 0.512
5000 0.019 0.513
6300 0.019 0.513
8000 0.020 0.513
10000 0.020 0.513
12500 0.023 0.513
16000 0.028 0.514
20000 0.034 0.514
25000 0.066 0.513
31500 0.138 0.513
40000 0.430 0.509
50000 0.950 0.496
N.B. Don't bother with the RS 1+1:6.45+6.45 as an input transformer, the matching between the two sections are not good.
Distortion goes up somewhat with low impedance headphones.
Best wishes,
Susan.
PMA said:
Thanks 🙂
I changed the transformers to use the Gardners 1:4 for input and an RS 228-393 1+1:2+2 for the output (wired the other way around as a 4:1 step down).
This is the performance into 600 ohms:
An externally hosted image should be here but it was not working when we last tested it.
Frequency THD% Vac
20.0 1.559 0.444
25.0 0.956 0.455
31.5 0.610 0.464
40.0 0.386 0.472
50.0 0.254 0.478
63.0 0.168 0.484
80.0 0.110 0.488
100 0.077 0.492
125 0.058 0.495
160 0.045 0.497
200 0.038 0.499
250 0.035 0.501
315 0.030 0.503
400 0.027 0.504
500 0.025 0.505
630 0.023 0.506
800 0.021 0.508
1000 0.020 0.508
1250 0.019 0.509
1600 0.019 0.511
2000 0.019 0.511
2500 0.019 0.511
3150 0.019 0.512
4000 0.019 0.512
5000 0.019 0.513
6300 0.019 0.513
8000 0.020 0.513
10000 0.020 0.513
12500 0.023 0.513
16000 0.028 0.514
20000 0.034 0.514
25000 0.066 0.513
31500 0.138 0.513
40000 0.430 0.509
50000 0.950 0.496
N.B. Don't bother with the RS 1+1:6.45+6.45 as an input transformer, the matching between the two sections are not good.
Distortion goes up somewhat with low impedance headphones.
Best wishes,
Susan.
Simple and tough
Yeah! I like this beefed up transistor radio solution, maybe it can be used as a guitar amplifier, if so it is possible to make a production easy, minimum part technology but... transformers are very expensive today, as you need at least three transformers it ends up in a very exclusive affair.
A transformer in the signal path is an obscure device, it is on the same list as capasitores, I think that the old audio engineer guru John Curl have talked about it in some interview some years ago.
But as a tough signal amplifier with minimal service needs over time and when it brokes it can be repaired very quick and cheap.
Mmm... maybe I will try one but I feel a bit ambivalent, Susan Parker what have you done to me.
I will build one mono amplifier from parts from an old Bouyer transistor amplfier with transformer output it shall be used as an guitar amplifier.
Kamskoma
Yeah! I like this beefed up transistor radio solution, maybe it can be used as a guitar amplifier, if so it is possible to make a production easy, minimum part technology but... transformers are very expensive today, as you need at least three transformers it ends up in a very exclusive affair.
A transformer in the signal path is an obscure device, it is on the same list as capasitores, I think that the old audio engineer guru John Curl have talked about it in some interview some years ago.
But as a tough signal amplifier with minimal service needs over time and when it brokes it can be repaired very quick and cheap.
Mmm... maybe I will try one but I feel a bit ambivalent, Susan Parker what have you done to me.
I will build one mono amplifier from parts from an old Bouyer transistor amplfier with transformer output it shall be used as an guitar amplifier.
Kamskoma
Re: Simple and tough
Hi Kamskoma,
Look forward to hearing of your progress 🙂
If you are concerned about cost for a guitar amp standard mains power toroids could be used - my test setup worked up to 25 kHz.
http://www.susan-parker.co.uk/zeus-toroid-amp-1.htm
How many watts are you looking at?
Best wishes,
Susan.
Hi Kamskoma,
kamskoma said:
Look forward to hearing of your progress 🙂
If you are concerned about cost for a guitar amp standard mains power toroids could be used - my test setup worked up to 25 kHz.
http://www.susan-parker.co.uk/zeus-toroid-amp-1.htm
How many watts are you looking at?
Best wishes,
Susan.
Zeus SE Line Preamp
Hi,
I have added some info on a single ended version of the 600/150 ohm line driver with THD v frequency comparison.
http://www.susan-parker.co.uk/zeus-line-driver-1.htm
This could use a single TX-102 magnetic volume control per channel instead of the Gardners input transformer.
Best wishes,
Susan.
Hi,
I have added some info on a single ended version of the 600/150 ohm line driver with THD v frequency comparison.
http://www.susan-parker.co.uk/zeus-line-driver-1.htm
An externally hosted image should be here but it was not working when we last tested it.
This could use a single TX-102 magnetic volume control per channel instead of the Gardners input transformer.
Best wishes,
Susan.
guitar and power
Hi Susan,
I think it will be Ok with ~15W a beginners electric guitar amplifier to begin with, I will try to put the guitar into a computers sound card and do the emulation in a software then from line out to the SP amplifier.
I have a line 600Ohm trasformatore from Lundahl ca it be used as input device or...
Kamskoma
Hi Susan,
I think it will be Ok with ~15W a beginners electric guitar amplifier to begin with, I will try to put the guitar into a computers sound card and do the emulation in a software then from line out to the SP amplifier.
I have a line 600Ohm trasformatore from Lundahl ca it be used as input device or...
Kamskoma
Re: guitar and power
Hi Kamskoma,
Okay, not a 2/3 horse amp then 🙂
Is the speaker impedance 8 or 16 ohms?
Probably. Which part type number is it so I can have a look at the spec?
Best wishes,
Susan.
Hi Kamskoma,
kamskoma said:
Okay, not a 2/3 horse amp then 🙂
Is the speaker impedance 8 or 16 ohms?
Probably. Which part type number is it so I can have a look at the spec?
Best wishes,
Susan.
Re: Simple and tough
Speaking of Lundahl, me thinks someone as Jeff Rowland will disagree.
kamskoma said:
Speaking of Lundahl, me thinks someone as Jeff Rowland will disagree.