Re: Almost done...
From what I can see in the photo, it looks like you've mounted the heatsinks with fins horizontally. I don't see how they would work anything close to optimal in that configuration, since convection wouldn't direct much airflow between the fins, and you're not making use of but a fraction of the surface area of the heatsinks.
syn08 said:
From what I can see in the photo, it looks like you've mounted the heatsinks with fins horizontally. I don't see how they would work anything close to optimal in that configuration, since convection wouldn't direct much airflow between the fins, and you're not making use of but a fraction of the surface area of the heatsinks.
Re: Re: Almost done...
Agreed but sometimes you have to work with what's available. Those HSs were around $50 (including shipping from the US) a pair on EBay. I was unable to find anything with vertical fins anywhere close to that price range. Ro-Theta here in Toronto quoted me $150 a piece (that is, $300/pair plus taxes).
Another issue was with the HS positions. Ideally they should be built as the two sides of the case. Well, I was unable to find a decent priced, right sized, case with integrated HSs that would not run into $500-$700, plus shipping from Europe or Thailand, plus taxes, plus customs, plus spares readily available. I am also not ready to invest in a metal shop so what was left as an option was a 4RU Bud Industries case for $150. Top cover is perforated, of course.
The HSs are oversized and the fans control (MAX6665ASA55) is taking care of any transient thermal conditions.
abzug said:
Agreed but sometimes you have to work with what's available. Those HSs were around $50 (including shipping from the US) a pair on EBay. I was unable to find anything with vertical fins anywhere close to that price range. Ro-Theta here in Toronto quoted me $150 a piece (that is, $300/pair plus taxes).
Another issue was with the HS positions. Ideally they should be built as the two sides of the case. Well, I was unable to find a decent priced, right sized, case with integrated HSs that would not run into $500-$700, plus shipping from Europe or Thailand, plus taxes, plus customs, plus spares readily available. I am also not ready to invest in a metal shop so what was left as an option was a 4RU Bud Industries case for $150. Top cover is perforated, of course.
The HSs are oversized and the fans control (MAX6665ASA55) is taking care of any transient thermal conditions.
Re: Re: Re: Re: Re: Re: Cmcl
The input impedance is rather low and the gain is 1. The target S/N for 1V RMS input is -120dB in an unweighted 20 khz bandwidth, but only measurements on the final PCB will show the exact values (measuring low noise levels on a breadboard is impossible).
Overall, the S/N ratio with the balanced input stage will be worse than the single ended input S/N. Transformers can potentially do better in terms of preserving the S/N and CMRR, but I don't really like the idea.
G.Kleinschmidt said:
The input impedance is rather low and the gain is 1. The target S/N for 1V RMS input is -120dB in an unweighted 20 khz bandwidth, but only measurements on the final PCB will show the exact values (measuring low noise levels on a breadboard is impossible).
Overall, the S/N ratio with the balanced input stage will be worse than the single ended input S/N. Transformers can potentially do better in terms of preserving the S/N and CMRR, but I don't really like the idea.
Re: Re: Re: Re: Re: Re: Re: Cmcl
What I did with my design was to make a pair of discrete opamps configured as unity gain followers / buffers for each input, to provide a high input impedance (100k). The noise contribution of the input followers can be made extremely small, and lower value feedback resistors are then allowed in the fully differential section.
Of couse, this adds a great deal more complexity, but that's all part of the fun of it, IMO.
Another question - My amp is a bridged design (two idependent, but identical power amps per channel), so I make use of both complementary outputs from the differential opamp stage.
How do you make use the differential outputs to drive the single ended input of the PGP?
One output must be used in reference to the other (you basically get OL linearity if you measure/take an output individually, referenced to ground).
Cheers,
Glen
syn08 said:
What I did with my design was to make a pair of discrete opamps configured as unity gain followers / buffers for each input, to provide a high input impedance (100k). The noise contribution of the input followers can be made extremely small, and lower value feedback resistors are then allowed in the fully differential section.
Of couse, this adds a great deal more complexity, but that's all part of the fun of it, IMO.
Another question - My amp is a bridged design (two idependent, but identical power amps per channel), so I make use of both complementary outputs from the differential opamp stage.
How do you make use the differential outputs to drive the single ended input of the PGP?
One output must be used in reference to the other (you basically get OL linearity if you measure/take an output individually, referenced to ground).
Cheers,
Glen
Re: Re: Re: Re: Re: Re: Re: Re: Cmcl
With two input buffers, have you estimated the CMRR? Worst case (opamp dispersion, resistor dispersion) its not going to be great...
The output is single ended, that's the output stage that I added to bruno's differential circuit. Full schematic will be posted on the web site as soon as I'm done with the measurements.
G.Kleinschmidt said:
With two input buffers, have you estimated the CMRR? Worst case (opamp dispersion, resistor dispersion) its not going to be great...
The output is single ended, that's the output stage that I added to bruno's differential circuit. Full schematic will be posted on the web site as soon as I'm done with the measurements.
Re: Re: Re: Re: Re: Re: Re: Re: Re: Cmcl
The CMRR really isn't effected much - input buffer opamps work fine for professional instrumentation amplifiers, afterall. The buffer opamps have 100% nfb, so (feedback) resistor dispertion doesn't come into it. Each opamp is identical and has a unity gain frequency >4MHz, so OLG throughout the entire audio spectrum and beyond is ample. Unity gain accuracy, phase shift deviation, Z-out, etc is therefore very tight.
syn08 said:
The CMRR really isn't effected much - input buffer opamps work fine for professional instrumentation amplifiers, afterall. The buffer opamps have 100% nfb, so (feedback) resistor dispertion doesn't come into it. Each opamp is identical and has a unity gain frequency >4MHz, so OLG throughout the entire audio spectrum and beyond is ample. Unity gain accuracy, phase shift deviation, Z-out, etc is therefore very tight.
Re: Re: Re: Re: Re: Re: Re: Cmcl
Ah yes - that was a nice thread over there in Power Supplies.
I've been looking for a solution that uses variable phase control on the triac. I wish I could figure out what's going on in the Bryston soft start circuit! I guess I'll start out with a simple circuit first and work my way up to a phase control circuit later.
syn08 said:
Ah yes - that was a nice thread over there in Power Supplies.
I've been looking for a solution that uses variable phase control on the triac. I wish I could figure out what's going on in the Bryston soft start circuit! I guess I'll start out with a simple circuit first and work my way up to a phase control circuit later.
Re: Re: Re: Re: Re: Re: Re: Re: Cmcl
A variable phase control circuit should not be a problem if a specialized ic is used. I was looking myself for such a solution and the problem is in how to power the controller. Due to the supply current requirements I doubt a capacitive divider can be used, while a power resistor is to me not a good option.
andy_c said:
A variable phase control circuit should not be a problem if a specialized ic is used. I was looking myself for such a solution and the problem is in how to power the controller. Due to the supply current requirements I doubt a capacitive divider can be used, while a power resistor is to me not a good option.
Re: phase control circuit
You mean due to DC (from asymmetrical chopping of the positive and negative half cycles)? Wouldn't this tend to go away once the 100 percent duty cycle is reached? Or is it some other cause you're referring to? It would be nice to get rid of the high-power resistor.
I did notice that the Bryston circuits have DC compensation that can buck out two diode drops (the full wave bridge with the large capacitors across it in the schematics). Are the Brystons known to have the buzzing transformer problem?
dimitri said:
You mean due to DC (from asymmetrical chopping of the positive and negative half cycles)? Wouldn't this tend to go away once the 100 percent duty cycle is reached? Or is it some other cause you're referring to? It would be nice to get rid of the high-power resistor.
I did notice that the Bryston circuits have DC compensation that can buck out two diode drops (the full wave bridge with the large capacitors across it in the schematics). Are the Brystons known to have the buzzing transformer problem?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.