Last night I was looking through data sheets and not finding anything that was really close. If I got the Rds to be as low then the Cin was much higher. I found some possibilities in new surface mounts from Vishay but haven't had time to do more today. I'm hoping tomorrow I can spend a couple of hours on it. I'll post with what I find if something looks good. By the way, this is helping me. Just being forced to go though data sheets is at least making me familiar with the names of parameters people consider important. It's a start anyway. Thanks
Rds is not important, it is not a switching application. Concentrate on capacitances. Also, it have to be available in the package you can use.
This one will fit, for example:
http://www.fairchildsemi.com/ds/FQ/FQD8P10.pdf
Also, this one should be widely available and cheap: http://www.datasheetcatalog.org/datasheet/irf/irf9610.pdf
This one will fit, for example:
http://www.fairchildsemi.com/ds/FQ/FQD8P10.pdf
Also, this one should be widely available and cheap: http://www.datasheetcatalog.org/datasheet/irf/irf9610.pdf
Also, look at forward transconductance speaking of input capacitance.
If transconductance of IRF9610 is 900 mA /Volt, transconductance of ZVP4105 is 50 mA/Volt, that means for the same change of drain current IRF9610 needs 18 times less voltage swing. It is equal to 18 times less of input capacitance, for the same time to drive by the same current. 170pF/18 = less than 10 pF: more than 4 times less than ZVP4105. However, for such conclusions we need to calculate according to expected working currents and voltages since both Ciss and gfs are not constant.
Output capacitance is also a significant factor because it is practically in emitter of the transistor that loads the tube, and I have concern about that. But it's too late, let's continue tomorrow.
If transconductance of IRF9610 is 900 mA /Volt, transconductance of ZVP4105 is 50 mA/Volt, that means for the same change of drain current IRF9610 needs 18 times less voltage swing. It is equal to 18 times less of input capacitance, for the same time to drive by the same current. 170pF/18 = less than 10 pF: more than 4 times less than ZVP4105. However, for such conclusions we need to calculate according to expected working currents and voltages since both Ciss and gfs are not constant.
Output capacitance is also a significant factor because it is practically in emitter of the transistor that loads the tube, and I have concern about that. But it's too late, let's continue tomorrow.
Hearinspace said:
Yes, you can stop on this one.
One problem with this gyrator is, it is self-destructive if to ground it's output suddenly. Adding a resistor in emitter you protect it from over-current, also output capacitance of the P-FET will play lesser role on high frequencies.
100 Ohm would be fine. about 100 mA peak, that means 30W dissipation on the output transistor shortly when it happens. During normal operation about 2V will drop on it, so a P-FET will still have anough.
I'll redraw a schemo later.
Hi Wavebourn,
OK , one more question as I'm getting ready to pick up the parts for this. Is there a way of doing this without the zeners? In part, I wonder about putting a part that is known for its noise in a high res. circuit. but also and more interesting to me, if the driver stage Vp (in the direct coupled circuit) can be set up to track appropriately with changes in the output stage Vpk due to mains voltage fluctuation it would seem to be kind of elegant. Is there any reason not to use resistors instead of the zener string?
thanks again!!
OK , one more question as I'm getting ready to pick up the parts for this. Is there a way of doing this without the zeners? In part, I wonder about putting a part that is known for its noise in a high res. circuit. but also and more interesting to me, if the driver stage Vp (in the direct coupled circuit) can be set up to track appropriately with changes in the output stage Vpk due to mains voltage fluctuation it would seem to be kind of elegant. Is there any reason not to use resistors instead of the zener string?
thanks again!!
You can do that.
For example, a string of 3 resistors for average voltage drops: about 12V in the middle + about 350V on bottom, the rest on the top. However, current through it should be significant against base current and it's variations of your BJT in cascode. Now, you may calculate power dissipated on such a string.
Yes, Zeners generate noises. Resistors generate noises as well.
You may find data about noise voltages of that Zeners, and use simple equations of their level on anode, keeping in mind a dynamic resistance of your current source and internal anode resistance of your tube. For better rejection of power supply ripples you will need to shunt your resistors by capacitors. But the same capacitors may be used to shunt Zeners that loose much less power and ensure more stable working points.
Speaking of an elegance, can you explain what do you mean?
I see absolutely no elegance in tracking of mains voltage fluctuations, and believe my power amps that don't depend on mains voltage fluctuations and need special very expensive power cords the same like cows need saddles, are much more elegant.
I've asked you already, "Do you want DC coupled stages?" Because if you do, I would draw and calculate you very similar triode load, but with feedback by DC from cathode of output tube, instead of anode of the driver.
For example, a string of 3 resistors for average voltage drops: about 12V in the middle + about 350V on bottom, the rest on the top. However, current through it should be significant against base current and it's variations of your BJT in cascode. Now, you may calculate power dissipated on such a string.
Yes, Zeners generate noises. Resistors generate noises as well.
You may find data about noise voltages of that Zeners, and use simple equations of their level on anode, keeping in mind a dynamic resistance of your current source and internal anode resistance of your tube. For better rejection of power supply ripples you will need to shunt your resistors by capacitors. But the same capacitors may be used to shunt Zeners that loose much less power and ensure more stable working points.
Speaking of an elegance, can you explain what do you mean?
I see absolutely no elegance in tracking of mains voltage fluctuations, and believe my power amps that don't depend on mains voltage fluctuations and need special very expensive power cords the same like cows need saddles, are much more elegant.
I've asked you already, "Do you want DC coupled stages?" Because if you do, I would draw and calculate you very similar triode load, but with feedback by DC from cathode of output tube, instead of anode of the driver.
I'll answer your last question first. Yes I will want to try the amp direct coupled but it's already built (breadboard) as a very straight forward cap coupled amp with simple cascoded Mosfet CCS loads on both stages. It sounds very good so far and it makes sense to me to proceed making one change at a time (so I can make sense of what gets what results). So after trying your SVCS in the cap coupled circuit , then I can rebuild for direct coupled.
As for the bit about elegance, of course if the whole amp is regulated it's another matter, but when the output stage is not regulated (as in the present breadboard amp) having a regulated driver means the driver's Vp is only "relevant" at one point and mismatched or arbitrary at all others. I just like the idea that everything tracks together in a circuit - ie . that all the voltages expand and contract in proportion so that at very least the output tube is kept from over conduction due to bad bias, and in a best case scenario both tubes run along anticipated/desirable load lines regardless of (realistic) mains fluctuations. That in my mind is very simple elegance , no matter that the intention behind it might not be obvious just from looking at the schematic.
I'm happy to try this both with zeners AND resistors , bypassed and not. My intention here is not to get you to do the final design of my amp for me for free. It really is to learn something about how to look at these things. Even a couple of days ago just being prodded to look through data sheets with the successive suggestions from you was a big help in forcing me to become familiar with some of the relevant parameters. I probably wouldn't have done that without your help/prodding, so it's really appreciated.
As for the bit about elegance, of course if the whole amp is regulated it's another matter, but when the output stage is not regulated (as in the present breadboard amp) having a regulated driver means the driver's Vp is only "relevant" at one point and mismatched or arbitrary at all others. I just like the idea that everything tracks together in a circuit - ie . that all the voltages expand and contract in proportion so that at very least the output tube is kept from over conduction due to bad bias, and in a best case scenario both tubes run along anticipated/desirable load lines regardless of (realistic) mains fluctuations. That in my mind is very simple elegance , no matter that the intention behind it might not be obvious just from looking at the schematic.
I'm happy to try this both with zeners AND resistors , bypassed and not. My intention here is not to get you to do the final design of my amp for me for free. It really is to learn something about how to look at these things. Even a couple of days ago just being prodded to look through data sheets with the successive suggestions from you was a big help in forcing me to become familiar with some of the relevant parameters. I probably wouldn't have done that without your help/prodding, so it's really appreciated.
Hearinspace; what kind of track do you mean? If due to starvation of output tube you get more distorted sound from it, you want your preamp to distort as well?
May be it is a good principle in friendship, when one friend suffers the other true friend have to co-suffer, but in electronics if a preamp can be kept in regimes without adding distortions regardless of what happens to an output stage, it has to be done.
But what is even better, to keep low distortions regardless of usage conditions; direct coupling and regime stabilization is a good way to go.
For the capacitive coupling, here is the current version of my sneaky load:
http://wavebourn.com/forum/download.php?id=125&f=7
May be it is a good principle in friendship, when one friend suffers the other true friend have to co-suffer, but in electronics if a preamp can be kept in regimes without adding distortions regardless of what happens to an output stage, it has to be done.
But what is even better, to keep low distortions regardless of usage conditions; direct coupling and regime stabilization is a good way to go.
For the capacitive coupling, here is the current version of my sneaky load:
http://wavebourn.com/forum/download.php?id=125&f=7
What kind of track? I'm not smart enough to say yet, so you win. . . . No , wait! maybe not ! Why does one have to suffer? and the other co-suffer? I like the idea of an unregulated circuit that keeps both triodes running on anticipated and desirable load lines wherever the mains voltage is (within normal tolerances of course). It would mean that the DCR of the entire circuit including the power supply would have to be designed to allow for it. (Surely engineers from pre-chip regulator days did this for circuits thought to warrant the extra effort/cost) If using choke filtering the chokes would most likely all have to be custom wound. This was one of the ideas behind the Axiom - a circuit I have never heard but thought seemed a neat idea after reading the designer's notes.
Anyway OK, I'll get the parts for this rev in a few days. I'll let you know as soon as it's in circuit
Anyway OK, I'll get the parts for this rev in a few days. I'll let you know as soon as it's in circuit
It is a regulated power supply that keeps both triodes running on anticipated and desirable load lines wherever the mains voltage is floating. For saving of costs the final one may run from unregulated supply. It is a trade-off. 60 years ago such trade-off was reasonable, except in instrumentation. That's why you don't see regulated power supplies in a vintage consumer gear, the exception is, VR tubes in some military super-heterodyne receivers to stabilize heterodyne frequencies. In some expensive audio amplifiers screen grid supplies were regulated.
Hi Wavebourn,
It's been a busy couple of weeks but I managed to put together one board (one part at a time) and tonight one channel is running with the SVCS. Settling with about a five volt rise in about as many minutes it's now sitting stable and sounds good, maybe really really good - easily smoother or less aggressive than the casoded Mosfet CCS it takes the place of. Of course both sides will be needed to really do a fair comparison.
I'll try to finish another one on the weekend.
Thanks !
It's been a busy couple of weeks but I managed to put together one board (one part at a time) and tonight one channel is running with the SVCS. Settling with about a five volt rise in about as many minutes it's now sitting stable and sounds good, maybe really really good - easily smoother or less aggressive than the casoded Mosfet CCS it takes the place of. Of course both sides will be needed to really do a fair comparison.
I'll try to finish another one on the weekend.
Thanks !
Just reporting in . . . .
Have two channels running now. The sound is so unexpected that I don't quite know what to think.
With the gain lower than with the previous CCS , and less dynamic and softer, the sound is a step in the opposite direction from what I've been working to cultivate, supposedly a negative result, yet the more I listen the more I like it. It's a little strange in that relief or definition between instruments is not as distinct, yet low level detail like room decay is more audible and much cleaner - I would have thought both qualities would be governed by the same performance characteristics .
Comparing the two plate loads, the SVCS seems to be netting a sound better suited for longer term semi-background music. (The previous circuit sounds more "live", but how long can you have a band wailing away in your living room before asking them to take a break?) They've been playing for about 6 hours .
I'll leave them for a few more days and then try a few things - a different PNP or even replacing with a Mosfet might tell me something about how sensitive the circuit is to parts choices. (I did find the 5731A to be an improvement over the MJE350 but better doesn't necessarily mean good.)
Any strong recommendations for best sounding PNP and P-Channel devices to try?
Thanks!
Have two channels running now. The sound is so unexpected that I don't quite know what to think.
With the gain lower than with the previous CCS , and less dynamic and softer, the sound is a step in the opposite direction from what I've been working to cultivate, supposedly a negative result, yet the more I listen the more I like it. It's a little strange in that relief or definition between instruments is not as distinct, yet low level detail like room decay is more audible and much cleaner - I would have thought both qualities would be governed by the same performance characteristics .
Comparing the two plate loads, the SVCS seems to be netting a sound better suited for longer term semi-background music. (The previous circuit sounds more "live", but how long can you have a band wailing away in your living room before asking them to take a break?) They've been playing for about 6 hours .
I'll leave them for a few more days and then try a few things - a different PNP or even replacing with a Mosfet might tell me something about how sensitive the circuit is to parts choices. (I did find the 5731A to be an improvement over the MJE350 but better doesn't necessarily mean good.)
Any strong recommendations for best sounding PNP and P-Channel devices to try?
Thanks!
Wavebourn SVCS, more questions
Hi Wavebourn,
Well , the new loads have been running a few days now and I think I love them. It's hard to know sometimes as it is doing something with the sound that I don't think I have ever heard before and can't quite figure out how to put a finger on it, but as I listen I like it more and more . . . . . . . . and more!
So, can we keep going? I'd like to learn more from you if I can and am wondering if you feel like giving any more guidance. I'd like to try doing something with the idea you mentioned (quoted above). If you have the time for it.
Well, they all seem to have thought it didn't need other ideas. I have been looking for alternative PNP transitors to replace the MJE5731 but haven't found anything yet. I'm don't know about MosFets. Would they be worth trying?
Thanks Wavebourn !
Wavebourn said:
Hi Wavebourn,
Well , the new loads have been running a few days now and I think I love them. It's hard to know sometimes as it is doing something with the sound that I don't think I have ever heard before and can't quite figure out how to put a finger on it, but as I listen I like it more and more . . . . . . . . and more!
So, can we keep going? I'd like to learn more from you if I can and am wondering if you feel like giving any more guidance. I'd like to try doing something with the idea you mentioned (quoted above). If you have the time for it.
Wavebourn said:
Well, they all seem to have thought it didn't need other ideas. I have been looking for alternative PNP transitors to replace the MJE5731 but haven't found anything yet. I'm don't know about MosFets. Would they be worth trying?
Thanks Wavebourn !
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.