A few corrections are in order :
1) If you use the single ended version with +/-24V rails with 1.3A bias, you might consider using 0R47 for R11 & R12, though 0R22 would also work.
2) You will not get 4.2V across R3 & R4. More like 3V.
3) R5-R8 are definitely incorrect and need changing. R5 & R7 want to be both 100R, R6 & R8 both 150R.
4) I would use 100R or 220R for R13 & R14.
But the simplest way to do this is to use solution A in post #8720.
In that case everything in the original schematics stays the same, except one additional 5 ohm resistor at the source of 2SJ74.
The latter now has an Idss of 0.8mA more than 2SK170.
Of course this entire thing only makes sense if you are using 2SK1530 / 2SJ201 as output devices.
Patrick
1) If you use the single ended version with +/-24V rails with 1.3A bias, you might consider using 0R47 for R11 & R12, though 0R22 would also work.
2) You will not get 4.2V across R3 & R4. More like 3V.
3) R5-R8 are definitely incorrect and need changing. R5 & R7 want to be both 100R, R6 & R8 both 150R.
4) I would use 100R or 220R for R13 & R14.
But the simplest way to do this is to use solution A in post #8720.
In that case everything in the original schematics stays the same, except one additional 5 ohm resistor at the source of 2SJ74.
The latter now has an Idss of 0.8mA more than 2SK170.
Of course this entire thing only makes sense if you are using 2SK1530 / 2SJ201 as output devices.
Patrick
Last edited:
Sch. changed to reflect the simplest solution offered by Patrick's degeneration trick. Tea Bag, I hope you don't mind.
The boards I had burned acomadated some machined pins so that the P and N-channel could be "degenerated" differently by paralleling in values and, indeed, you could tweak the performance a bit -- and NP had mentioned somewhere ab initio that you could tweak the 0.47R drain resistors a bit to fiddle with the performance.
As the P and N JFETs have different capacitance one can also use different gate stoppers --
> Of which is sounds like option b in 8720 sounds easiest, swap 10 for 15R on R2, sounds easy enough. I will put it into the queue.
Please read carefully :
R5 & R7 want to be both 100R, R6 & R8 both 150R.
The 2SJ74 now has an Idss of 0.8mA more than 2SK170.
The schematics in #8743 is correct, except that voltage across R3 / R4 is 2.7V when using 0R22 for R11, R12.
Gate resistors still too low, for my taste.
Patrick
Please read carefully :
R5 & R7 want to be both 100R, R6 & R8 both 150R.
The 2SJ74 now has an Idss of 0.8mA more than 2SK170.
The schematics in #8743 is correct, except that voltage across R3 / R4 is 2.7V when using 0R22 for R11, R12.
Gate resistors still too low, for my taste.
Patrick
Last edited:
Newbie Rectifier question - finishing my F5
I finished my F5 last night. I'm starting to mount components in their final places. I have two questions I could find the answers to:
Do the rectifiers need to be on a big heatsink (the ones for the IRF MOSFETs), or just a piece of aluminum plate?
Do they need to be electrically insulated from their heatsink?
I finished my F5 last night. I'm starting to mount components in their final places. I have two questions I could find the answers to:
Do the rectifiers need to be on a big heatsink (the ones for the IRF MOSFETs), or just a piece of aluminum plate?
Do they need to be electrically insulated from their heatsink?
the bridge rectifiers do not need electrical insulation. A little thermal compound helps.
To220 diodes are connected to the thermal interface. Careful you don't accidentally short them.
Most rectifiers do not need sinks for extra cooling.
ClassA rectifiers do however dissipate a lot of heat continuously. A small heatsink is usually sufficient. A case floor is good enough for most installations.
But that increases the in chassis Ta.
To220 diodes are connected to the thermal interface. Careful you don't accidentally short them.
Most rectifiers do not need sinks for extra cooling.
ClassA rectifiers do however dissipate a lot of heat continuously. A small heatsink is usually sufficient. A case floor is good enough for most installations.
But that increases the in chassis Ta.
If you are using Peter's diode boards, (or others) with MUR 3060 diodes the answer is YES, they need the isolation pads under them. This is sort of an exception and surprised me, I already had them tightened down with thermal grease when I happened to re read one of Peter's posts about it. The MUR 3060 definitely needs to be isolated. My case is made from 1/8 inch thick aluminum ( 6061 ) plate, and the diodes are just bolted to the floor panel.
Russellc
I'm using 625-GBPC3508-E4 from Mouser - Vishay Bridge Rectifiers 35 Amp 800 Volt in a GBPC case.
Now I have to swap a few more resistors.
What would increasing the MOSFET gate stoppers do for SQ or performance?
Currently I am using .47R Source resistors
An externally hosted image should be here but it was not working when we last tested it.
I have tried solution A using this method, and it work!
Another Newbie question
As I said in an earlier post, I'm finishing up my F5, it's a stock build. This is the first electronic item I've ever built - it was surprisingly easy.
I jury-rigged the new F5 build to hear if it was working. I put one channel together and ran the input to my computer sound out. The speaker was the closest one handy - a satellite from a Logitech Z-680.
Everything worked, and it was quite satisfying to hear music from something I put together, but...
The gain was not at all what I expected. I don't know anything about audiophile equipment, but I could only drive the speaker to moderate sound levels.
Would I benefit from a preamp? I don't know how a computer output stacks up against a preamp output, are they similar, or will a preamp get one more volume in the end?
Is the sound level just the innate limitation of the F5?
I know folks will say cascode and get more output devices, and I guess I could, I'm hoping there's a simpler answer.
As I said in an earlier post, I'm finishing up my F5, it's a stock build. This is the first electronic item I've ever built - it was surprisingly easy.
I jury-rigged the new F5 build to hear if it was working. I put one channel together and ran the input to my computer sound out. The speaker was the closest one handy - a satellite from a Logitech Z-680.
Everything worked, and it was quite satisfying to hear music from something I put together, but...
The gain was not at all what I expected. I don't know anything about audiophile equipment, but I could only drive the speaker to moderate sound levels.
Would I benefit from a preamp? I don't know how a computer output stacks up against a preamp output, are they similar, or will a preamp get one more volume in the end?
Is the sound level just the innate limitation of the F5?
I know folks will say cascode and get more output devices, and I guess I could, I'm hoping there's a simpler answer.
No, there is no simple answer.
F5 has a gain of 6, compared to most power amps which has a gain of 10 to 30.
This is one reason why I am not using the F5 in my current system.
If you have a preamp of a gain of 10, then you should have enough gain to do some volume control.
More output devices and higher voltages will give you more power, not more gain.
Patrick
F5 has a gain of 6, compared to most power amps which has a gain of 10 to 30.
This is one reason why I am not using the F5 in my current system.
If you have a preamp of a gain of 10, then you should have enough gain to do some volume control.
More output devices and higher voltages will give you more power, not more gain.
Patrick
The F5 has low gain. Changing the feedback resistor to 100ohm (if you use two, omitting one of them) doubles the gain and is able to reach maximum voltage swing (of the amplifier's capability, IIRC 14V) with some tradeoffs in distortion.
If you are satisfied with the sound quality, using a preamp will create the additional swing required to reach full output. Soundcards usually have very low output levels to start with, unless they are professional interfaces (most of which will deliver voltage output required for the F5 to reach maximum swing).
If you are satisfied with the sound quality, using a preamp will create the additional swing required to reach full output. Soundcards usually have very low output levels to start with, unless they are professional interfaces (most of which will deliver voltage output required for the F5 to reach maximum swing).
Changing the feedback resistor will change both the level of feedback and the close loop bandwidth. So it will sound different.
How different ? I have not tried, so don't know.
Even with a full 2Vrms output from a CD player, gain of 6 is still not full swing.
Going balanced will give you a factor of 2 more.
But I presume you still want say 10-20dB attenuation for volume control.
So you cannot avoid another gain stage, I am afraid.
Of course you can also solve them problem by using 100dB speakers, like horns, etc.
Then you get the 20dB back, and can live with the low output of the F5.
Patrick
How different ? I have not tried, so don't know.
Even with a full 2Vrms output from a CD player, gain of 6 is still not full swing.
Going balanced will give you a factor of 2 more.
But I presume you still want say 10-20dB attenuation for volume control.
So you cannot avoid another gain stage, I am afraid.
Of course you can also solve them problem by using 100dB speakers, like horns, etc.
Then you get the 20dB back, and can live with the low output of the F5.
Patrick
Patrick,
Recall, this is my first try with high end equipment, I'm not sure I'm following all your thoughts.
You're right, I'd like to keep volume control - I'd planned on an integrated Lightspeed, but that's out for now.
What is the practical outcome of a preamp vs. a balanced F5 (I assume you're talking about the circuit you posted)? How would going balanced give me a 2x in gain compared to a standard F5?
What Pass (or DIY) preamps might fit the bill?
Is going balanced a better answer than bridging?
Recall, this is my first try with high end equipment, I'm not sure I'm following all your thoughts.
You're right, I'd like to keep volume control - I'd planned on an integrated Lightspeed, but that's out for now.
What is the practical outcome of a preamp vs. a balanced F5 (I assume you're talking about the circuit you posted)? How would going balanced give me a 2x in gain compared to a standard F5?
What Pass (or DIY) preamps might fit the bill?
Is going balanced a better answer than bridging?