Mean you this device from the Italian brand "GRAAF" ?
GM 70 OTL Preamplificatore phono
But I think rather, you mean the GM20 about
http://www.diyaudio.com/forums/tubes-valves/142528-graaf-gm20-mods-help.html
Attachments
Last edited:
SIT/SST I-V characteristics
The 'SIT' is a vertical channel JFET with 'triode-like' IV dependence. By contrst, MOSFETs (DMOS, LDMOS, etc.) and BJTs, as well as MESFETs exhibit saturated, 'penthode-like' IV charctersitics.
For eons, HiFi enthisiasts bet their lives on the the more 'antural' sound coming out from triode (tubes) amplifiers than from 'transistors' (or 'solid state') amplifiers.
I poseted here that, althogh my ears are not good enough, I did devlop lots of SIT=SST (Solid State Triode) power devices, some for audio amps.
We did not see the market there and abandoned the audio apps. project some 10 years back.
I warms my heart to see some interest in SSTs still exist and I will be glad to post more info ir needed.
May you all get all the watts you need,
AIC.
The 'SIT' is a vertical channel JFET with 'triode-like' IV dependence. By contrst, MOSFETs (DMOS, LDMOS, etc.) and BJTs, as well as MESFETs exhibit saturated, 'penthode-like' IV charctersitics.
For eons, HiFi enthisiasts bet their lives on the the more 'antural' sound coming out from triode (tubes) amplifiers than from 'transistors' (or 'solid state') amplifiers.
I poseted here that, althogh my ears are not good enough, I did devlop lots of SIT=SST (Solid State Triode) power devices, some for audio amps.
We did not see the market there and abandoned the audio apps. project some 10 years back.
I warms my heart to see some interest in SSTs still exist and I will be glad to post more info ir needed.
May you all get all the watts you need,
AIC.
I did not open thread about output caps, since findings here are almost empirical. And not only for caps, choice of schematics, of power supplies variants, of parts, are also supported by mainly audio expertise. Unfortunately, after completion of sound tests of dozens of schematics and design approaches, I did not find any real support from widely accepted theoretical postulates (THD, spectra etc.). I also played with genuine SITs, and would not overestimate possible magic sound effect from their use.
One of my findings, is that reasons of losses of sound microdetails can not be revealed by routine measurements or simulation software. In particular, the PSRR and NFB concepts, while approved by simulations, do not work good enough for preserving microdetails in audio signal. The same is valid for electrolytic caps (especially at low-current stages), for use of Zeners, for integrated voltage regulators, for electro-magnetic interference. "Green light" for using of NFB is given by high intrinsic linearity and high speed of stages, especially of output stage of an amp, and SITs and jFETs are good for this.
Comparing coomon source vs common drain at the output stage, I got confirmations in favour of ciuffolli's solution, again it is mostly empirical experience. Shunt power supplies for other non-ciuffolli stages also have crucial positive effects for sound.
Non-linear gate capacitances of FETs cause problems, and the only way to make transistors really competitive with tubes - using RF parts, with much smaller and more linear capacitances. The RF jFETs are the best to me at present.
Having new Nelson's SITs at hand, I would simply try to put them at the output of this schematics, with higher current driver stage http://www.diyaudio.com/forums/solid-state/197493-jfet-amp-current-nfb.html
Last edited:
more on "SITs": output and feedback caps vs. voltage while used in common source
FWIW: The (very few) commercilly available power (Silicon) SITs are 'normally on' or 'depletion' type device, like the GaAs MESFETs and most low power Si JFETs. One unique characteristics to the SIT is that, as the drain-to source channel is depleted, it operates with high internal electric fields.
Another result of the fully depleted D-S 'channel' and the high internal fields is the relatively low Cgd and Cgs dependence upon Vds, both under low and high current channel currents.
One have to remember that the channel current (electrons) is 'pulled over' the internal potential barrier formed under the source, in the gate region. This is waht was referred to by Nishizawa as (Elctro)Static (Filed) Induction and hence the SIT name.
In part, this barrier dependent, depleted channel features, give the SIT a higher linearity (or, indirectly, a low THD coeff. for audio folks folks) when compared to BJTs and MOFETs.
The resulting nonsaturated ('triode like') IV charcteristics make it behave (and hence 'sound') like the so much cherished decades old a(and still going) vacuum tubes.
Conclusion: SITs are extremelly well suited for applicathins requireing superior linearity or low distortions.
As I mentioned, the HF Power Gain is low, but this should not be a problem when using SITs in audio amp apps. below, say, 1MHz.
I hope this helps someone...
May you all get (very) low ditortions of your input (signals)!
AIC.
FWIW: The (very few) commercilly available power (Silicon) SITs are 'normally on' or 'depletion' type device, like the GaAs MESFETs and most low power Si JFETs. One unique characteristics to the SIT is that, as the drain-to source channel is depleted, it operates with high internal electric fields.
Another result of the fully depleted D-S 'channel' and the high internal fields is the relatively low Cgd and Cgs dependence upon Vds, both under low and high current channel currents.
One have to remember that the channel current (electrons) is 'pulled over' the internal potential barrier formed under the source, in the gate region. This is waht was referred to by Nishizawa as (Elctro)Static (Filed) Induction and hence the SIT name.
In part, this barrier dependent, depleted channel features, give the SIT a higher linearity (or, indirectly, a low THD coeff. for audio folks folks) when compared to BJTs and MOFETs.
The resulting nonsaturated ('triode like') IV charcteristics make it behave (and hence 'sound') like the so much cherished decades old a(and still going) vacuum tubes.
Conclusion: SITs are extremelly well suited for applicathins requireing superior linearity or low distortions.
As I mentioned, the HF Power Gain is low, but this should not be a problem when using SITs in audio amp apps. below, say, 1MHz.
I hope this helps someone...
May you all get (very) low ditortions of your input (signals)!
AIC.
More about SITs.
The 'generic's' SIT IV characteristics are identical to those of a vacuum triode, that is they exhibit either and exponential (SIT) or a power function dependence of the drain current (SST). They are referred to as 'short channel IV characteristics, unlike the saturated 'long channel' IVs of a HV MOSFET. This is in
One may see a decent rendition of IV charactristics of a generic SIT deigned for audio apps. in US Pat. 5648664.
Keep in mind that the SITs suitable for quality audio amps are 'normally on' devices and they need a negative gate voltage to be turned off.
All power SITS (scarcely) available since the late 1970's, are Silicon devices. Recently, SiC SITs have become available. They are not actually 'pure' barrier field control devices as the 'real' SITs are supposed to be. As Vertical Channel (Junction) FETs (VFETs), they may do a good job in power amps but I'm not sure about their 'intrinsic' linearity and I suspect that some negative feedback or feed forward ckts. may be needed.
Low THD to all,
AIC.
The 'generic's' SIT IV characteristics are identical to those of a vacuum triode, that is they exhibit either and exponential (SIT) or a power function dependence of the drain current (SST). They are referred to as 'short channel IV characteristics, unlike the saturated 'long channel' IVs of a HV MOSFET. This is in
One may see a decent rendition of IV charactristics of a generic SIT deigned for audio apps. in US Pat. 5648664.
Keep in mind that the SITs suitable for quality audio amps are 'normally on' devices and they need a negative gate voltage to be turned off.
All power SITS (scarcely) available since the late 1970's, are Silicon devices. Recently, SiC SITs have become available. They are not actually 'pure' barrier field control devices as the 'real' SITs are supposed to be. As Vertical Channel (Junction) FETs (VFETs), they may do a good job in power amps but I'm not sure about their 'intrinsic' linearity and I suspect that some negative feedback or feed forward ckts. may be needed.
Low THD to all,
AIC.
perhaps this has been already addressed elsewhere (and if so, feel free to re-direct me), but if i may ask (and if this is an OK place to ask), how simple would it be to adopt the NewWatt prototype circuit discussed in the SixMoons article to run on 2SK77's? i don't understand enough about transistor characteristics to eyeball that on my own. i see that they were mentioned at the end of the article in the context of higher power output. i happen to very much enjoy both my Yamaha B-1 and my DIYed F5, but i just have to wonder what it would be like to have somewhat of a hybrid!
Last edited:
I think these devices would be ideal for the amp designed by Susan Parker.
http://www.diyaudio.com/forums/solid-state/42259-zero-feedback-impedance-amplifiers.html
Al the voltage amplification is done by the transformers, the Mosfets are used for Current amplification. Sounds like the SIT's NP is using would be a great fit.
http://www.diyaudio.com/forums/solid-state/42259-zero-feedback-impedance-amplifiers.html
Al the voltage amplification is done by the transformers, the Mosfets are used for Current amplification. Sounds like the SIT's NP is using would be a great fit.
Make new SITs, optimized for audio apps?
The 2SK77 or2SK183 are ‘antiques’, fabricated using an exotic 'buried gate epitaxial process'. The MwT SITs were ‘true’ planar devices and it will be a breeze for any power IC foundry to make device at much lower cost, higher uniformity and higher yield that the old devices I see mentioned on thi s site.
Assuming enough devices are needed, it may make sense to ‘ask’ MwT Inc., now part of IXYS, to consider dusting off the old SIT/SST info and get back to the market with an audio optimized FET?
One should benefit from 21st century’s technologies and I'm not sure the SiC SITs are better than MOSFEts whan used for quality audio amps!
I’m sure that, say, a 300V, 200mA SIT in a TO247 or similar, should cost much (much!) less than the $60/ea I noticed mentioned on this site.
Can anybody supply a list of must have device characteristics, such as Gfs, BV (this is a blocking not a breakdown voltage), Vgs swing, Cgs, Cgd, etc.?
Be careful what (transitory) you wish, you may get it!
AIC.
The 2SK77 or2SK183 are ‘antiques’, fabricated using an exotic 'buried gate epitaxial process'. The MwT SITs were ‘true’ planar devices and it will be a breeze for any power IC foundry to make device at much lower cost, higher uniformity and higher yield that the old devices I see mentioned on thi s site.
Assuming enough devices are needed, it may make sense to ‘ask’ MwT Inc., now part of IXYS, to consider dusting off the old SIT/SST info and get back to the market with an audio optimized FET?
One should benefit from 21st century’s technologies and I'm not sure the SiC SITs are better than MOSFEts whan used for quality audio amps!
I’m sure that, say, a 300V, 200mA SIT in a TO247 or similar, should cost much (much!) less than the $60/ea I noticed mentioned on this site.
Can anybody supply a list of must have device characteristics, such as Gfs, BV (this is a blocking not a breakdown voltage), Vgs swing, Cgs, Cgd, etc.?
Be careful what (transitory) you wish, you may get it!
AIC.
i only mentioned the 2SK77 because i have 4 good ones sitting in my B-1. i figured that if i heard enough good things about these new SIT amps, one day i might get bored and...i think you see where this is going. i wasn't talking about a new production run.
perhaps about
http://www.diyaudio.com/forums/soli...to3-like-yamaha-2sk77-nec-2d363-overview.html
are the wanted URLs.
This is nice, I only saw Japanese data so far.
Does your B-I OM also gives data for the other V-FET devices ?
My German B-I OM doesn't say anything about technical data of the transistors ...
The data clearly shows, why Yamaha put a 2SK75 in front of the 2SK77:
The non linear capacitances Cgd and Cgs are extremely high and need a strong driver to cope with the transit frequency modulation of the output device (and the resultant PIM distortion).
Modern SiC SIT are certainly much better here, but the junction caps could be also much more non linear (relatively), because a more complex and dense process geometry can provide dramatic capacitance changes at lower voltages (Which is the result of developing transistors of lower gate charge and high switching efficiencies).
I would not put a 2SK77 into a Zen like circuit, because the highly nonlinear feedback capacitance would appear parallel to the feedback resistor, which is certainly a bad idea, if such resistance is not smaller than a few tens of Ohms.
And: The best place for a 2SK77 is a B-I 😛
Does your B-I OM also gives data for the other V-FET devices ?
My German B-I OM doesn't say anything about technical data of the transistors ...
The data clearly shows, why Yamaha put a 2SK75 in front of the 2SK77:
The non linear capacitances Cgd and Cgs are extremely high and need a strong driver to cope with the transit frequency modulation of the output device (and the resultant PIM distortion).
Modern SiC SIT are certainly much better here, but the junction caps could be also much more non linear (relatively), because a more complex and dense process geometry can provide dramatic capacitance changes at lower voltages (Which is the result of developing transistors of lower gate charge and high switching efficiencies).
I would not put a 2SK77 into a Zen like circuit, because the highly nonlinear feedback capacitance would appear parallel to the feedback resistor, which is certainly a bad idea, if such resistance is not smaller than a few tens of Ohms.
And: The best place for a 2SK77 is a B-I 😛
perhaps about
Yamaha B-1 owner roll-call (and info repository) - AudioKarma.org Home Audio Stereo Discussion Forums
and
Sound Thinking -> Yamaha B-1 in the House
are to find this informations.
And an auction:
http://www.canuckaudiomart.com/deta...power_amp_with_uc1_meters__ru1_remote_adaptor
Yamaha B-1 owner roll-call (and info repository) - AudioKarma.org Home Audio Stereo Discussion Forums
and
Sound Thinking -> Yamaha B-1 in the House
are to find this informations.
And an auction:
http://www.canuckaudiomart.com/deta...power_amp_with_uc1_meters__ru1_remote_adaptor
Last edited:
ha, forgot about starting that AK thread!
Jon, realistically i'm not going to rip out the 2SK77's from my B-I. if it wasn't in good cosmetic shape and wasn't totally refurbished, it might be a different story though!
i don't want this discussion to veer off-track, though. i have one statement as to why the 2SK77 would not adapt well to the new SIT prototype circuit, are others in agreement?
Jon, realistically i'm not going to rip out the 2SK77's from my B-I. if it wasn't in good cosmetic shape and wasn't totally refurbished, it might be a different story though!
i don't want this discussion to veer off-track, though. i have one statement as to why the 2SK77 would not adapt well to the new SIT prototype circuit, are others in agreement?
Last edited:
I have some 2SK77B's, and they are very nice parts, particularly in terms
of the dissipation figures, but they would not be appropriate to the kind
of single-stage amplifiers that I am currently playing with. The input
capacitance is about 10 nF for a single-ended Class A which means that
they have to be driven by a low source impedance, lower than you expect
in audio preamps.
As I indicated previously, I have an article due shortly which addresses some
of these issues for several parts.
😎
of the dissipation figures, but they would not be appropriate to the kind
of single-stage amplifiers that I am currently playing with. The input
capacitance is about 10 nF for a single-ended Class A which means that
they have to be driven by a low source impedance, lower than you expect
in audio preamps.
As I indicated previously, I have an article due shortly which addresses some
of these issues for several parts.
😎
- Status
- Not open for further replies.