Terry,
Don't worry about having one heat sink (side?) smaller than the other. With 200W per channel, you'll need all the heat sinking you can accomodate.
If it's rated 200 W /ch, and you'll use it for music, I would probably count on 100W sinking ability per channel. Yes, class B is inefficient, but then again, most of the time you won't uss more than 20% of the power available if you want good head-room for music transients. Short transients will be handled by the heat capacity of the heat sink material. Long term power will be handled by the heat sink thermal rating.
Anyway... considering that there will be thermal impedance in the transistor casings and mounting insulation (don't forget that part!) I would not recomment a heat sink temperature above 50deg. Celcius. Given a 25C room temperature, you'll need to sink 100W with a 25degree temperature rise. Thus, your heat sinks should have at least 0.25K/W rating or better (meaning: lower value) per channel.
This will easily fill the entire side of your amp. (and yes, they're a convenient way of making the sides of the enclosure).
In conclusion, your 2- or 3ch. considerations may be decided simply by the heat sinking capability of your enclosure, and not by the fear of complexity of the PSU.
If the amp will be placed in a location with good air circulation at all sides, you may consider using either the front or rear of the enclosure as location for the 3rd channel heat sink area.
On another note:
I'm glad if you find my notes useful. I think it's the first time (I'm aware of) that the document has come to use as a whole. I doubt that I have time at the moment to change it, but let me know if you come across serious errors or things that need improvement.
Jennice
Don't worry about having one heat sink (side?) smaller than the other. With 200W per channel, you'll need all the heat sinking you can accomodate.
If it's rated 200 W /ch, and you'll use it for music, I would probably count on 100W sinking ability per channel. Yes, class B is inefficient, but then again, most of the time you won't uss more than 20% of the power available if you want good head-room for music transients. Short transients will be handled by the heat capacity of the heat sink material. Long term power will be handled by the heat sink thermal rating.
Anyway... considering that there will be thermal impedance in the transistor casings and mounting insulation (don't forget that part!) I would not recomment a heat sink temperature above 50deg. Celcius. Given a 25C room temperature, you'll need to sink 100W with a 25degree temperature rise. Thus, your heat sinks should have at least 0.25K/W rating or better (meaning: lower value) per channel.
This will easily fill the entire side of your amp. (and yes, they're a convenient way of making the sides of the enclosure).
In conclusion, your 2- or 3ch. considerations may be decided simply by the heat sinking capability of your enclosure, and not by the fear of complexity of the PSU.
If the amp will be placed in a location with good air circulation at all sides, you may consider using either the front or rear of the enclosure as location for the 3rd channel heat sink area.
On another note:
I'm glad if you find my notes useful. I think it's the first time (I'm aware of) that the document has come to use as a whole. I doubt that I have time at the moment to change it, but let me know if you come across serious errors or things that need improvement.
Jennice
Hey Terry,
Ive just finsihed building 2 p101 boards for my amp prject, im runing them monoblock with 2 transformers like so:
if anyone can comment on that as a layout , would be GREATLY appreciated, (mostly in ref to keeping the 0.1r resistors with only 4 caps)
Terry, if ur having any probs and need me to compare anything with my boards to yours etc (voltagesmeasurements, etc ) , pls dont hesitate to ask, my boards are up and running on my test (single) power supply , you shouldnt have any dramas tho.
mine are the 2 FEt versions at the moment due to budget, ill be adding the extra 2 fets later on.
do you have Rev a boards that require the mod? , make sure u dont forget it if so
goodluck with all.
Ive just finsihed building 2 p101 boards for my amp prject, im runing them monoblock with 2 transformers like so:
if anyone can comment on that as a layout , would be GREATLY appreciated, (mostly in ref to keeping the 0.1r resistors with only 4 caps)
Terry, if ur having any probs and need me to compare anything with my boards to yours etc (voltagesmeasurements, etc ) , pls dont hesitate to ask, my boards are up and running on my test (single) power supply , you shouldnt have any dramas tho.
mine are the 2 FEt versions at the moment due to budget, ill be adding the extra 2 fets later on.
do you have Rev a boards that require the mod? , make sure u dont forget it if so
goodluck with all.
mAJORD,
Are you using 3 caps per rail or per transformer ?
If you're using 4 per transformer, I wouldn't suggest the resistor. (unless you're using massive caps).
Regarding the fuse, I would consider to move it before the switch.
This way, swich arching accidents and possibe failure will be less dramatic.
Also, have you considered a soft start circuit? It'll be easier on all your PSU components, and you can rate the fuse nearer it's intended value, and still prevent it from blowing when turning the PSU on.
Jennice
Are you using 3 caps per rail or per transformer ?
If you're using 4 per transformer, I wouldn't suggest the resistor. (unless you're using massive caps).
Regarding the fuse, I would consider to move it before the switch.
This way, swich arching accidents and possibe failure will be less dramatic.
Also, have you considered a soft start circuit? It'll be easier on all your PSU components, and you can rate the fuse nearer it's intended value, and still prevent it from blowing when turning the PSU on.
Jennice
Hi mAJORD,
Thanks for the reply and the offer. I will be taking you up on that.
I wasn't planning on using two trannies unless I use all three PCBs. Rod Elliot seemed to think one was sufficient for stereo. O showed somewhere earlier in this thread how he suggested I hook up two trannies for the three channel amp. To go mono block I would need three trannies to go three channel. I don't want to get that aggressive yet.
I do have the Revision A boards. They came with the traces already cut and I just had to add the jumpers.
I went ahead and ordered all twelve MOSFETs from Tech-DIY. Cost me $102 including shipping. They were all matched.
May I ask what rail voltages you are using with your trannies?
Thanks
Hi Jens,
I have built a soft start circuit for my mp, but have been told I may not need it due to my trannies being IE as opposed to Toroid. What do you think?
As far as the heat sinks go, if I'm reading correctly, with the Conrad heatsinks I will need a 12" X 4" X 2" heatsink for each PCB. I had better stick to stereo if that is the case.I don't think I want to have a much bigger case than that. I don't really like the idea of running a third heatsink across the front or back. I was hoping to mount two boards on one side and one on the other. I didn't realize I would need that much heatsing for three channels.
Blessings, TErry
Thanks for the reply and the offer. I will be taking you up on that.
I wasn't planning on using two trannies unless I use all three PCBs. Rod Elliot seemed to think one was sufficient for stereo. O showed somewhere earlier in this thread how he suggested I hook up two trannies for the three channel amp. To go mono block I would need three trannies to go three channel. I don't want to get that aggressive yet.
I do have the Revision A boards. They came with the traces already cut and I just had to add the jumpers.
I went ahead and ordered all twelve MOSFETs from Tech-DIY. Cost me $102 including shipping. They were all matched.
May I ask what rail voltages you are using with your trannies?
Thanks
Hi Jens,
I have built a soft start circuit for my mp, but have been told I may not need it due to my trannies being IE as opposed to Toroid. What do you think?
As far as the heat sinks go, if I'm reading correctly, with the Conrad heatsinks I will need a 12" X 4" X 2" heatsink for each PCB. I had better stick to stereo if that is the case.I don't think I want to have a much bigger case than that. I don't really like the idea of running a third heatsink across the front or back. I was hoping to mount two boards on one side and one on the other. I didn't realize I would need that much heatsing for three channels.
Blessings, TErry
Hi Terry,
Could you give a name/type of the Conrad heat sink you have in mind?
As for the soft-start, I would surely use it.
Unless your transformers are much smaler than I recall, your switch, your capacitors & bridge-diodes, as well as your mains fuse will have a tough life ( = shorter life).
Your mains switch will be easy/cheap to replace, but the caps won't.
How much capacity did you plan on using?
Jennice
Could you give a name/type of the Conrad heat sink you have in mind?
As for the soft-start, I would surely use it.
Unless your transformers are much smaler than I recall, your switch, your capacitors & bridge-diodes, as well as your mains fuse will have a tough life ( = shorter life).
Your mains switch will be easy/cheap to replace, but the caps won't.
How much capacity did you plan on using?
Jennice
Hi Jennice,
To get to the rating you suggest I would need the MF30-151.5. You can see it here.
Looks like they will cost $50 each.
My transformers are 64 volt rails @ 400mV.
I plan to use These caps for this amp. I had panned to use 6 or 8 of them but am open for suggestions.
Thanks, Terry
To get to the rating you suggest I would need the MF30-151.5. You can see it here.
Looks like they will cost $50 each.
My transformers are 64 volt rails @ 400mV.
I plan to use These caps for this amp. I had panned to use 6 or 8 of them but am open for suggestions.
Thanks, Terry
Hi Terry,
I don't understand this one:
>>My transformers are 64 volt rails @ 400mV.<<
anyway, with any transformer suited for a 2 x 200W amplifier, I would definitely use a softstart circuit. Especially with the kind of capacity you'll be using. I would use at least 4 caps per rail (8 in total).
As for the heat sinks: Yes, they're outrageously expensive
and make me feel like a 
in the wrong business.
The ones in my previous amp project (the one I use for playing music these days) are 420mm wide, 200 mm high and 42mm deep, and rated 0.25 K/W. Either mine are rated conservatively, or yours are optimistic, but I agree on your choice. They look huge, but make excellent sides of your amp. With right-angle aluminum profiles, you'll be making yourself a sturdy amp frame/enclosure, with reasonable overall cost.
Jennice
I don't understand this one:
>>My transformers are 64 volt rails @ 400mV.<<
anyway, with any transformer suited for a 2 x 200W amplifier, I would definitely use a softstart circuit. Especially with the kind of capacity you'll be using. I would use at least 4 caps per rail (8 in total).
As for the heat sinks: Yes, they're outrageously expensive
and make me feel like a in the wrong business. The ones in my previous amp project (the one I use for playing music these days) are 420mm wide, 200 mm high and 42mm deep, and rated 0.25 K/W. Either mine are rated conservatively, or yours are optimistic, but I agree on your choice. They look huge, but make excellent sides of your amp. With right-angle aluminum profiles, you'll be making yourself a sturdy amp frame/enclosure, with reasonable overall cost.
Jennice
Terry,
Have you seen this section on their heat sink calculations?
http://www.conradheatsinks.com/technical_details.html
This may explain why your smaller heat sink is rated better than mine!
The problem with their specification is that you do absolutely not want a 80 deg. Celcius heat sink temperature in such a project!
Thus, your heat sinking ability will be reduced somewhat, but you should still be able to get a decent result. Heat sinking and capacitors are two things that one can never get too much of.
Jennice
Have you seen this section on their heat sink calculations?
http://www.conradheatsinks.com/technical_details.html
This may explain why your smaller heat sink is rated better than mine!
The problem with their specification is that you do absolutely not want a 80 deg. Celcius heat sink temperature in such a project!
Thus, your heat sinking ability will be reduced somewhat, but you should still be able to get a decent result. Heat sinking and capacitors are two things that one can never get too much of.
Jennice
Terry,
Just another note about Conrad Heatsinks (and no I don't work for them) $50 seems to be alot of money but when you factor in how much it is going to cost you to build the sides anyway, and believe me when I say the finish of these heatsinks is nice enough that it would be a shame not to use them as the sides. They are very high quality and I wouldn't be affraid to shell out twice as much for them. Hopefully they don't read this
I have to buy seven of them.
I haven't completly read the notes Jen put on here, buy the way nice work, thanks, but looking at the construction notes Rod has, he suggests a heatsink with a minimum of 0.6deg C/W, with this I was going to go with the MF20-75 here http://www.conradheatsinks.com/products/flat100_350.html#MF25. I think I better go and read Jen's notes some more.
Just another note about Conrad Heatsinks (and no I don't work for them) $50 seems to be alot of money but when you factor in how much it is going to cost you to build the sides anyway, and believe me when I say the finish of these heatsinks is nice enough that it would be a shame not to use them as the sides. They are very high quality and I wouldn't be affraid to shell out twice as much for them. Hopefully they don't read this
I have to buy seven of them. I haven't completly read the notes Jen put on here, buy the way nice work, thanks, but looking at the construction notes Rod has, he suggests a heatsink with a minimum of 0.6deg C/W, with this I was going to go with the MF20-75 here http://www.conradheatsinks.com/products/flat100_350.html#MF25. I think I better go and read Jen's notes some more.
Terry:
MY Transformers are 300Va each, 40-0-40 , giving + - 56v rails, (approx) , thisis the max voltage for a 2 fet version.
Jennice:
4 caps per transformer, (note ive chopped off 2 of the 6 in the orignal single tranny version)
so i figure thats 2 per RAIL, they are 4000uF each, 70v 4.6amp ripple rating.
MY Transformers are 300Va each, 40-0-40 , giving + - 56v rails, (approx) , thisis the max voltage for a 2 fet version.
Jennice:
4 caps per transformer, (note ive chopped off 2 of the 6 in the orignal single tranny version)
so i figure thats 2 per RAIL, they are 4000uF each, 70v 4.6amp ripple rating.
mAJORD,
I wouldn't use resistors with only 2 caps per rail. Also, I think 2 x 4000uF is not a lot for such an amp.
Audiofng,
why seven heat sinks (such an odd number) ???
Rod's note on 0.6K/W is what I would consider a minimum, and depending on supply voltage. When running with few power devices for a given rating, you will have higher thermal losses, which gives higher operting temperture on chip. (Again, see my notes.) Therefore, I suggest a bigger heat sink, especially given the manufacturers own technical notes (link in a previous post), which in reality tells the user that their thermal ratings are not applicable for projects like power amplifiers.
Jennice
I wouldn't use resistors with only 2 caps per rail. Also, I think 2 x 4000uF is not a lot for such an amp.
Audiofng,
why seven heat sinks (such an odd number) ???
Rod's note on 0.6K/W is what I would consider a minimum, and depending on supply voltage. When running with few power devices for a given rating, you will have higher thermal losses, which gives higher operting temperture on chip. (Again, see my notes.) Therefore, I suggest a bigger heat sink, especially given the manufacturers own technical notes (link in a previous post), which in reality tells the user that their thermal ratings are not applicable for projects like power amplifiers.
Jennice
Jennice, thanks for the advice,
i can use more caps if need be, i have 10 in total at my disposal, the only reason i went for what you see there is that the ESP suggested psu has 6x 4700uf Caps for 2 channels , im using 4x 4000 for 1 channel,
do you or anyone have a formula relevent to AB amps where I can take into account things like losses in the transformer etc etc? for a given load asnd power output, i found it difficult at the time of buying to find sturdy info on selecting capacitor values!
about the only thing i learnt for sure was more lower value caps the better, to lower ESR.
cheers!
i can use more caps if need be, i have 10 in total at my disposal, the only reason i went for what you see there is that the ESP suggested psu has 6x 4700uf Caps for 2 channels , im using 4x 4000 for 1 channel,
do you or anyone have a formula relevent to AB amps where I can take into account things like losses in the transformer etc etc? for a given load asnd power output, i found it difficult at the time of buying to find sturdy info on selecting capacitor values!
about the only thing i learnt for sure was more lower value caps the better, to lower ESR.
cheers!
mAJORD,
Hmm.... Maybe I've started to slide over the edge, off the deep end, over the years...
I'm the kinda' guy who can't get the PSU bulky enough. (Within my reasonable limits.
Thus, I'm using a 550 VA transformer (the highest rating - except torroidal, which didn't fit my budget when custom made - that could fit my enclosure) with 50.000uF per rail for a 2 x 150W amplifier. Obviously this needs a soft-start circuit...
Transformer losses, or output impedance, or load regulation (different ways to describe the same phenomenon) should be stated by the manufacturer. Alternatively, you can load the transformer (without rectifier) wih different resistive loads, and make a V-out vs. I_load plot. From this, you can calculate the output resistance.
Regarding capacitors, it's a subjective thing. A circuit (the amplifier PCB/design) with really good Power Supply Rejection Ratio (often mentioned as PSRR in data sheets) needs less supply smoothing. (it is less affected by ripple in the supply, caused from capacitor voltage drops between the transformer voltage peaks). IC op-amps are among the best circuits (as far as I have seen) to achieve high PSRR. Discrete designs, especially those which are not symmetrical in design) are more likely to be affected == lower PSRR.
Especially in power amps, the objective is to minimise voltage drops due to output load current. This is achieved by lowering the output resistance of the PSU (and amplifier design!), which for the PSU is done with capacitors. The more capacity, the less voltage drop during audio transients which can require high currents. There are mathematical rules for calculating voltage drop per time unit for a given capacitor and load, but I have not seen anything specific for audio. Basically the same rules, except that your current needs vary with the music signal, and are therefore not constant. This can be inserted into integrals (do you call the mathematical formula so, also? (possible language barrier here!)), but I don't think it's worth the time.
Jennice
Hmm.... Maybe I've started to slide over the edge, off the deep end, over the years...
I'm the kinda' guy who can't get the PSU bulky enough. (Within my reasonable limits.
Thus, I'm using a 550 VA transformer (the highest rating - except torroidal, which didn't fit my budget when custom made - that could fit my enclosure) with 50.000uF per rail for a 2 x 150W amplifier. Obviously this needs a soft-start circuit...
Transformer losses, or output impedance, or load regulation (different ways to describe the same phenomenon) should be stated by the manufacturer. Alternatively, you can load the transformer (without rectifier) wih different resistive loads, and make a V-out vs. I_load plot. From this, you can calculate the output resistance.
Regarding capacitors, it's a subjective thing. A circuit (the amplifier PCB/design) with really good Power Supply Rejection Ratio (often mentioned as PSRR in data sheets) needs less supply smoothing. (it is less affected by ripple in the supply, caused from capacitor voltage drops between the transformer voltage peaks). IC op-amps are among the best circuits (as far as I have seen) to achieve high PSRR. Discrete designs, especially those which are not symmetrical in design) are more likely to be affected == lower PSRR.
Especially in power amps, the objective is to minimise voltage drops due to output load current. This is achieved by lowering the output resistance of the PSU (and amplifier design!), which for the PSU is done with capacitors. The more capacity, the less voltage drop during audio transients which can require high currents. There are mathematical rules for calculating voltage drop per time unit for a given capacitor and load, but I have not seen anything specific for audio. Basically the same rules, except that your current needs vary with the music signal, and are therefore not constant. This can be inserted into integrals (do you call the mathematical formula so, also? (possible language barrier here!)), but I don't think it's worth the time.
Jennice
hehe, Yes Id like to run more filtering, but of course $ becomes an issue then, the capacitors would end up costing more than the rest of the amplifer, and I dont see wheather that expence will be worth it in the end, how much of a realistic advantage would me using 50000uF of filtering per rail , vs 8000, bearing in mind this is a 2 fet version atm, probably pushing about 100watt output into 8ohm , and the more important question is how much extra over what I have now WOULD be worth the $,
cheers
Jimmy
cheers
Jimmy
Well I sent an email to conrad heatsinks but they never wrote back. I went ahead and bought this heatsink figuring to cut it in half or there-abouts, for the two sides of my amp. Is there a way I can calculate how large I would need each heat sink to be? If I just cut into two pieces, it would make my amp 9 3/4" deep and 5 12" tall, that is just slightly larger than my Hafler P230. It's rated at 115wpc @ 8ohms.
The Heatsink I purchased appears to have a pretty thick backplate. I'm not sure if that helps or not.
So, tell me what you think. Will this be enough?
Thanks, Terry
The Heatsink I purchased appears to have a pretty thick backplate. I'm not sure if that helps or not.
So, tell me what you think. Will this be enough?
Thanks, Terry
I'm looking at a toroid transformer It says it is 62+62+17+17V 1000VA
This is the diagram it shows.
What kind of rail voltage should I expect from this?
Would this work in place of my two Hafler trannies?
Thanks, Terry
PS
This is what the test results show;
Voltage Test Result:
TEST CONDITION: Apply variable voltage to primary coil (115V terminals) from 100-180VAC. No load on secondary coils.
1. Primary V = 100VAC, Primary I = .04A
2. Primary V = 110VAC, Primary I = .05A
3. Primary V = 120VAC, Primary I = .06A
4. Primary V = 130VAC, Primary I = .07A
5. Primary V = 140VAC, Primary I = .10A
6. Primary V = 150VAC, Primary I = .13A
7. Primary V = 160VAC, Primary I = .30A
8. Primary V = 170VAC, Primary I = .95A
Dielectric Test:
TEST CONDITION: Apply dielectric meter on primary coil and secondary coil; and increase voltage up to 3500VAC. No initiate any spark.
Load Test Result:
TEST CONDITION: Input 120VAC 60Hz to the primary coil; parallel 2 secondary coils, 5 ohm 250W resistors as dummy load and room temperature at 25 degree C. Temperature rises 6 degree C after an hour test at load #5.
1. Primary I = 0.05A, Secondary V = 63.5VAC at 0.0A
2. Primary I = 2.24A, Secondary V = 62.3VAC at 4.2A
3. Primary I = 3.33A, Secondary V = 61.7VAC at 6.2A
4. Primary I = 6.40A, Secondary V = 59.9VAC at 12.0A
5. Primary I = 9.30A, Secondary V = 58.3VAC at 17.5A
Dimension – 5.9” Diameter x3.35” High
Weight – 15 LB net.
Wire lead – 12” color-coded.
Thanks
This is the diagram it shows.
An externally hosted image should be here but it was not working when we last tested it.
What kind of rail voltage should I expect from this?
Would this work in place of my two Hafler trannies?
Thanks, Terry
PS
This is what the test results show;
Voltage Test Result:
TEST CONDITION: Apply variable voltage to primary coil (115V terminals) from 100-180VAC. No load on secondary coils.
1. Primary V = 100VAC, Primary I = .04A
2. Primary V = 110VAC, Primary I = .05A
3. Primary V = 120VAC, Primary I = .06A
4. Primary V = 130VAC, Primary I = .07A
5. Primary V = 140VAC, Primary I = .10A
6. Primary V = 150VAC, Primary I = .13A
7. Primary V = 160VAC, Primary I = .30A
8. Primary V = 170VAC, Primary I = .95A
Dielectric Test:
TEST CONDITION: Apply dielectric meter on primary coil and secondary coil; and increase voltage up to 3500VAC. No initiate any spark.
Load Test Result:
TEST CONDITION: Input 120VAC 60Hz to the primary coil; parallel 2 secondary coils, 5 ohm 250W resistors as dummy load and room temperature at 25 degree C. Temperature rises 6 degree C after an hour test at load #5.
1. Primary I = 0.05A, Secondary V = 63.5VAC at 0.0A
2. Primary I = 2.24A, Secondary V = 62.3VAC at 4.2A
3. Primary I = 3.33A, Secondary V = 61.7VAC at 6.2A
4. Primary I = 6.40A, Secondary V = 59.9VAC at 12.0A
5. Primary I = 9.30A, Secondary V = 58.3VAC at 17.5A
Dimension – 5.9” Diameter x3.35” High
Weight – 15 LB net.
Wire lead – 12” color-coded.
Thanks
still4given said:
I know when my friend tried to email them they had some problems with there server. He was persistant and they finally got back to him. He got through and and the service was top notch.
I have looked at the type of heat shink you bought but decided agianst them as they are not finished (anodized or coated) and it is difficult to find them here locally. By the time I would get it and cut it and finish it. They cost more than Conrads. Although doing it your way is a little more DIY, and that is what this is all about.
Just my $.02 worth.
Can't wait to see you finish them and hear what your opinion of them is.
thanks
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.