| Kilowatt |
What does he use for output devices in this amp? http://www.vision.net.au/~anthony/1kwamp.htm
I believe whatever is used in this design would be the best option for me to use in my own big MOSFET amp design. He says they are IRF devices, but doesn't give a specific part number. I find this very frustrating. I would love to know more about that amp, but there is no information on it. That's rediculous. |
|
|
| djdan |
IRFP240 / IRFP9240 - 10 Pairs
You can see a little small amplifier (AV800), similar with the 1Kw, in same Site and you can buy the PCB . If want more power, you can add more IRFP, becouse it is very simple and make a supply with +/-100v. But if you do that , replace all electrolitic capacitors with other with same value but 160v rated.
Becouse IRFP devices are cheap you can run with 16 pairs and it will survive in any circumstances, anywhere , on stage or home, many years.
In my opinion , Mr. Holton make a great design and this amplifier is one of the best on WEB.
I will buy next month PCB for stereo N- Amplifier and AV-800.
New link for Mr Anthony Holton site:
http://www.aussieamplifiers.com/
Link for 800W amp. http://members.optusnet.com.au/~aussieamps/avamp.htm
Link for PCB ( looks great )http://members.optusnet.com.au/~aussieamps/ampkits.htm
Good luck Kilowatt ! |
|
|
| AudioFreak |
well Kilowatt lets see what we can deduce from the information he does give us.........
They are IRF devices ....
They are plastic TO-247.....
There is 20 output devices (10 N-channel and 10 P-channel).......
The supply is +/-110V DC
It is a push/pull topology
It has the same noise figures as the symetrical amp he makes therefore it is almost cirtainly the same class.... Class AB in this case....
it outputs 1050W into 4ohms which is which is about 23Amps peak.....
this limits it down to only a couple of suitable devices .... and IRFP240/IRFP9240 would seem to be the most likely suspects.
hope this helps...
also, anthony is a member of this forum so he may even tell you himself. |
|
|
| Kilowatt |
Thanks, that makes two of us that thinks they are probably IRFP240/IRFP9240 devices.
I would love to see the plans for that amp, so, Anthony, if you're listening, may I please have some information about it or plans? I'd appreciate it.
Thanks |
|
|
| Nelson Pass |
Golly, that looks a lot like the output stage to
my X150, rated at 150 watts. :) |
|
|
| WhtCastle |
| Drop that baby down a couple ohms and lets cook some food on that sucker :D |
|
|
| AudioFreak |
| quote: | Originally posted by Nelson Pass
Golly, that looks a lot like the output stage to
my X150, rated at 150 watts. :) |
Yeah i know :) but your X150 has a fair bit more bias current running thru it .... and anyway, you always like to go to extremes dont you :) |
|
|
| Kilowatt |
| Looks pretty puny compared to your X1000, doesn't it Neslon. :) |
|
|
| djk |
| "The supply is +/-110V DC " I wonder how well the IRFP9240 will hold up.When the amp goes to the rail the off side transistors have to stand up to 220V.And if the AC line goes 5% high you have 230V on the 200V part. |
|
|
| AudioFreak |
| I guess one can only hope that the rails will sag substantially... anyway ... its rated @ 550W into 8ohms which only needs about +/-93V............. The problem is that IRF dont make any very high power P-Channel Mosfets rated above -200V |
|
|
| Aussieamps |
Greetings Gentlemen
The 1kw amplifier does indeed use IRFP240 and IRFP9240 devices
The +-110 volt rails drop down to 96 volts under load.
If you wish to see the design for this amplifier you only need to visit my site and download the PDF file for the AV800 amplifier.
It is basicly the same amplifier except with fewer o/p devices.
The new web address is www.aussieamplifiers.com
Any other questions? just let me know I will do my best to answer them...
Warm regards
Anthony Holton
:) |
|
|
| AudioFreak |
| it wouldnt be such a bad idea to make 2 smaller amps of say 300W in 8ohms based on this amp then run them bridged into an 8ohm speaker.... you should get around 1200W into that 8ohm load and it would give a much bigger safety margin on the o/p devices..... i'd just run each amp with say 16 o/p devices (8 x IRFP240 and 8 x IRFP9240) with a big power supply. All said and done this should be enough given that this is a low bias Class AB design. |
|
|
| Kilowatt |
| It would if you plan on bridging the 1000W amps:D |
|
|
| Aussieamps |
Greetings once again...
To reply to Audiofreak.
When bridging any amplifier into say 8 Ohms, only double the power o/p is realised. In theory it should be 4 times the power o/p. but in practise because of power supply loss and amplifier inefficency it is only 2 times.
So a 300 watt into 8 Ohm amplifier when bridged will develop only 600 watts into 8 Ohms.
The 1kw and AV800 amplifier can easierly be bridged with no additional circuitry, because both amplifiers have balanced i/p stages. One only has to use the inverting i/p of one of the modules and the non-inverting i/p on the other then connect the load across the active o/p's of each amplifier module.
The AV800 Module when bridged produces 900 watts RMS into 8 Ohms and over 1600 watts RMS into 4 Ohms....
I have heard of no failures with this amplifier.
I hope this helps....
Warm regards
Anthony Holton
:) |
|
|
| AudioFreak |
| quote: | Originally posted by Aussieamps
Greetings once again...
To reply to Audiofreak.
When bridging any amplifier into say 8 Ohms, only double the power o/p is realised. In theory it should be 4 times the power o/p. but in practise because of power supply loss and amplifier inefficency it is only 2 times.
So a 300 watt into 8 Ohm amplifier when bridged will develop only 600 watts into 8 Ohms.
The 1kw and AV800 amplifier can easierly be bridged with no additional circuitry, because both amplifiers have balanced i/p stages. One only has to use the inverting i/p of one of the modules and the non-inverting i/p on the other then connect the load across the active o/p's of each amplifier module.
The AV800 Module when bridged produces 900 watts RMS into 8 Ohms and over 1600 watts RMS into 4 Ohms....
I have heard of no failures with this amplifier.
I hope this helps....
Warm regards
Anthony Holton
:) |
Well i cant agree with that sorry...
If you bridge 2 amps the voltage across the speaker should double and if you double the voltage you double the current which gives 4 times the power .... you've then just got to make sure that the power supply and the o/p's are upto the job. |
|
|
| Aussieamps |
Have you ever built an amplifier in bridge mode?
I can tell you I have built dozens of them and not one of them produced 4 times the output power, regardless of how big the power supply was or how many o/p devices it had.....
The AV800 amplifier has a 2KVA transformer with a 100,000uf bank of caps and it doesn't produce 4 times the power o/p in bridge mode...
:) |
|
|
| Kilowatt |
| I'll have to partially agree with AudioFreak. Two bridged OPS's will provide almost twice the power into twice the impedance as one normal OPS, because the voltage doubles. Theoretiacally 4 times the power into the same load, but that would be like hooking up each amp to half it's nominal recommended load, In which case, I would agree with Anthony that you would not get 4 times the power. We all know that amps do not produce twice the power into half the load. As far as two bridged AV800s with a sufficient power supply and OPS doing 1600W into 4 ohms and not 8, That doesn't seem right, is that a typo? All I know for sure is that bridged amps normally produce twice the power into twice the impedance. Why wouldn't they? They would each see the nominal load (half the load used when bridged). Maybe it's different with the AV800, I don't know. |
|
|
| AudioFreak |
| quote: | Originally posted by Kilowatt
Theoretiacally 4 times the power into the same load, but that would be like hooking up each amp to half it's nominal recommended load, In which case, I would agree with Anthony that you would not get 4 times the power. We all know that amps do not produce twice the power into half the load. As far as two bridged AV800s with a sufficient power supply and OPS doing 1600W into 4 ohms and not 8, That doesn't seem right, is that a typo? All I know for sure is that bridged amps normally produce twice the power into twice the impedance. Why wouldn't they? They would each see the nominal load (half the load used when bridged). Maybe it's different with the AV800, I don't know. [/B] |
Yeah i agree with what you have said .... most amps cant double there output into half the load but they can come very close!.... Anothony is proposing that amps produce the same power into half the load just with less voltage and more current ... this would give twice the power when bridged.... many many many amps do far better than this.
I also know many amps that when bridged can produce twice the power into twice the load but do not have the current capacity to supply double that into the original load...
A 2kVa transformer with 100,000uF is not that big for a 800W amp... |
|
|
| Aussieamps |
OK
Bridging works like this in theory.
You have an amplifier which has say a 2kw power supply.
Its power o/p into 8 Ohms is 100 watts RMS (28.4 VAC @ 3.55 Amperes RMS= 100.82 watts RMS)
If you bridge this same amplifier into an 8 Ohm load in theory you should get twice the voltage across the 8 Ohm load (56.8 VAC RMS @ 7.1 Amperes RMS = 403.28 Watts RMS)
Or 4 times the power o/p
If you then half the impedance or load on this bridged amplifier to say 4 Ohms. In theory you should get twice the o/p power again.
(56.8 VAC @ 14.2 amperes = 806.56 Watts RMS)
Which is 8 times the power o/p into 8 Ohms on an non-bridged amplifier or 4 times power o/p @ the same impedance of 4 Ohms.
Now this is assuming ideal amplifiers and power supplies with no losses what so ever
In practice this doesn't happen....
If you look at any commercial amplifier which operates in bridge mode, they only ever double their power o/p into any given load equal to the non bridged load...
My AV800 amplifier produces over 870 watts rms into 4 Ohms non bridged. In bridge mode it produce about 1600 watts rms into 4 Ohms with a 2KVA transformer. Into 8 Ohms bridged it produces about 870 watts. which is about double its power o/p into 8 Ohms of 450 watts RMS.
The only commercial amplifier that could possibly come close to 4 times the o/p power would be Crown amplifiers. Which use Grounded Bridge Technology and even then they a subject to power losses and produce 4 times the power o/p or greater by driving into impedances of 2 Ohms or less...
Another amplifier would be Krell and not even Nelson Pass amplifiers x1000 can produce 4 times the power o/p in bridge mode using just two single module amplifiers into 8 Ohms...
I am willing to be corrected here, But I just don't see it...
regards
Anthony Holton:) |
|
|
| Kilowatt |
AudioFreak:
Hmmm, he is, isn't he. That's of course just not true, unless half the power is wasted as heat, because we all know that an amplifier's OPS does not function like an output transformer in this manner (lowering the voltage, increasing the current). So, I wonder why the AV800 can not put out more than twice the power into the same load, even when bridged. Most amps definately can.
Anthony, were these bridged output ratings actually measured in bridge mode in the real world, or just assumed based on measurments taken from non-bridged operation? If so, why is it that bridged AV800's only do 900W into 8 ohms, 1600W into 4 ohms? |
|
|
| Aussieamps |
the figures quoted are result of real world measurements across the load.
The only explanation would be supply regulation, as the DC voltage on the supply caps drops to about 85 volts under load...
The way Crown achieve power o/p approching 4 times the power o/p is because they use tightly regulated SMPS... |
|
|
| AudioFreak |
| quote: | Originally posted by Aussieamps
OK
Bridging works like this in theory.
You have an amplifier which has say a 2kw power supply.
Its power o/p into 8 Ohms is 100 watts RMS (28.4 VAC @ 3.55 Amperes RMS= 100.82 watts RMS)
If you bridge this same amplifier into an 8 Ohm load in theory you should get twice the voltage across the 8 Ohm load (56.8 VAC RMS @ 7.1 Amperes RMS = 403.28 Watts RMS)
Or 4 times the power o/p
If you then half the impedance or load on this bridged amplifier to say 4 Ohms. In theory you should get twice the o/p power again.
(56.8 VAC @ 14.2 amperes = 806.56 Watts RMS)
Which is 8 times the power o/p into 8 Ohms on an non-bridged amplifier or 4 times power o/p @ the same impedance of 4 Ohms.
Now this is assuming ideal amplifiers and power supplies with no losses what so ever
In practice this doesn't happen....
If you look at any commercial amplifier which operates in bridge mode, they only ever double their power o/p into any given load equal to the non bridged load...
My AV800 amplifier produces over 870 watts rms into 4 Ohms non bridged. In bridge mode it produce about 1600 watts rms into 4 Ohms with a 2KVA transformer. Into 8 Ohms bridged it produces about 870 watts. which is about double its power o/p into 8 Ohms of 450 watts RMS.
The only commercial amplifier that could possibly come close to 4 times the o/p power would be Crown amplifiers. Which use Grounded Bridge Technology and even then they a subject to power losses and produce 4 times the power o/p or greater by driving into impedances of 2 Ohms or less...
Another amplifier would be Krell and not even Nelson Pass amplifiers x1000 can produce 4 times the power o/p in bridge mode using just two single module amplifiers into 8 Ohms...
I am willing to be corrected here, But I just don't see it...
regards
Anthony Holton:) |
I agree with the 1st part of that post but i would suggest you check your facts .... Pass X-1000 is a bridged monoblock before you start :) also 2 such units are indeed very capable of producing 4000W into 8ohms .... that is 4x the output power of a single X-1000 into the same 8ohm load :) also the professional yamaha amps produce twice the power into twice the load in bridged configuration but are unable to produce 4 times the output into the original load because the power supply is current limited. |
|
|
| Kilowatt |
Also, here's an example that proves that more than 2x the power is attainable using bridge mode (and it's not class D). This is the power ouput of Alpine's MRV-1507:
450W x 2 (4 Ohm Stereo) 1500 x 1 (Bridged 4 Ohm)
This is the max power, RMS is (they claim) 225W/900W respectively, which also proves the point, even if that 900W figure is a slight exaggeration.
Perhaps it is all in the power supply, an Anthony has suggested. |
|
|
| AudioFreak |
| I agree it is hugely dependent on the capabities of the power supply .... also most the specs for car amps are voltage limited not current limited .... mind you, if your bridging 2 amps to achieve 4 times the power then each of them must supply only double that of their normal output.... with a BIG power supply and plenty of o/p devices, this shouldnt be too much to ask.... after all, this is the equivalent of running the amp into half the impedance ... upto a point, this is quite doable. |
|
|
| michel |
Hi aussiamps,
If you use electronicsworkbench, and/or Multisim, i would like to send you a conceptual bridging schematic for your consideration...and comment?!? |
|
|
| AudioFreak |
| quote: | Originally posted by Aussieamps
My AV800 amplifier produces over 870 watts rms into 4 Ohms non bridged. In bridge mode it produce about 1600 watts rms into 4 Ohms with a 2KVA transformer. Into 8 Ohms bridged it produces about 870 watts. which is about double its power o/p into 8 Ohms of 450 watts RMS.
regards
Anthony Holton:) |
it appears we have misunderstood each other.... you say your AV800 can do 450W into 8ohms and 870W into 4ohms .... well if you bridge these amps into an 8ohm load then each will see the equivalent of 4ohms .... so each will produce 870W into the load ... giving a total of 1740W into the 8ohm load between the 2 amps .... this is almost 4 times the power of a single amp into 8ohms so it works quite nicely with theory ... sure i am willing to concede 30W loss per amp ... i just dont know how one of us didnt pick up on this earlier.... when bridging you'll have a minimum of 2 amps ...assuming 2 amps.... each produces the same voltage as they would have in non bridge config ... but because when one amp is + the other is - the voltage between them is double that of what a single amp would have produced on it's own.... so they are called upon to deliver twice the normal current ... which is the equivalent of running each amp into 1/2 the original load... now each amp has produced twice it's normal power but the voltage across the speaker terminals is doubled and so has the current so the speaker will get 4 times the power.... no bridging theory states that each single amp should have to produce 4 times the power into the same load .... only twice the power .... it is the total power into the speaker that goes up 4 fold compared to a single amp so if you have an amp rated for 2 x 100W into 8ohms, 2 x 200W into 4ohms then it should be rated 1 x 400W into 8ohms bridged ... sure, in practice it'll be a little less but that about does it.
So i stand by my earlier post ....
| quote: | Originally posted by AudioFreak
it wouldnt be such a bad idea to make 2 smaller amps of say 300W in 8ohms based on this amp then run them bridged into an 8ohm speaker.... you should get around 1200W into that 8ohm load and it would give a much bigger safety margin on the o/p devices..... i'd just run each amp with say 16 o/p devices (8 x IRFP240 and 8 x IRFP9240) with a big power supply. All said and done this should be enough given that this is a low bias Class AB design. |
in the above example, each amp is rated @ 300W into 8ohms so will give probably just less than 600W into 4ohms....now you have 2 such amps... if you bridge these into an 8ohm speaker then each will effectively see 4ohms and so produce just under 600W ... now each amp is making 600W into that load so there is 1200W into the load.... just like i originally stated it. |
|
|
| Kilowatt |
| I thought we were talking about the total power of the two amps all along. I guess not. |
|
|
| AudioFreak |
| Well kilowatt.... you and i were but it appears Anthony was not. |
|
|
| Kilowatt |
| Any idea how big of a compressor I would need to cool a 7200W class B MOSFET amp as an alternative to huge heatsinks? |
|
|
| AudioFreak |
| Kilowatt..... clearly quality is not the main concern here as Class B has rather large distortion figures..... you could also try Class G which is more efficient than Class AB but should sound better than Class B |
|
|
| Kilowatt |
Well, class AB, whatever. I think perhaps Randy Sloan has a different opinion on what class AB means.
Quality is of very great concern, I am hoping for distortion figures less than 0.1%. |
|
|
| PassFan |
Kilowatt:
Randy Sloan is knowlegable enough to understand the theory behind the amplifier classes as well as the rest of us. To form an opinion about something as basic a rule as this, and then put it out there for all to see would not be very smart. I would have to give Randy more credit than that. Perhaps you've taken him out of context. I don't think that finding a circuit with less than 0.1% distortion should be a problem. However, at 7200watts the problem will be in a large enough power supply to drive it without distortion and how to move it. I see on another thread you may build seperate amps at 900 watts a channel. These supplies would be a more realistic goal. Good luck. |
|
|
| Kilowatt |
The design of this amp will be based very much on what Randy Sloan has written, he's very knowledgeable, he calls it's class B, it has very low distortion (.0038%).
Anyway, the 7200W amp will have 8 x 900W channels. I want to cool it with a compressor like from an old freezer. I want to know how to figure out how big of a compressor I would need to achieve a certatin thermal rating for my heatsinks. |
|
|
| JoeBob |
| Well, just figure out how many watts of heat your amp will be putting out and then get ahold of a compressor rated to handle that amount of heat. Be warned though, freezer/fridge compressors aren't meant to handle a constant load of heat, they're meant to cool down a small amount of heat and then keep that object cold. You might be better off using a dehumidifyer compressor, they're more "powerfull" then fridge ones usually. But even then, I'm not sure how many watts they can handle, I just know I've used one to freeze an AMD PC processor pretty well (but they're obviously not all the same). With large amounts of heat you'd be much better off going H2O cooling, and if it really is alot of heat (I don't know how much heat would get put off by a 7200W amp) you could easilly go with several sperate loops, say a heat exchanger, pump and radiator per channel or per 2 channels... |
|
|
| Kilowatt |
| A 7200W class B MOSFET amp would dissipate maybe 4000W of heat, probably not much more in a worst-case scenerio. Would that be too much for a refridgerator compressor? I'm thinking a water or antifreeze cooling system would be bulky compared to a compressor, and would not be able to lower the temperature as much. Is this true? |
|
|
| JoeBob |
Well, water would work better then antifreeze, but 60% H2O and 40% methanol in a closed loop system is a very good way to go if you don't want stuff to grow (I gather that's why you said antifreeze). And yes, it could be rather bulky, but if done well you could probably minimize space required. But 4000W is much too much for a freezer compressor, I've had a fridge compressor choke on 200W of heat. If you really want to go with the compressor route you'd need lots of little freon compressors (maybe one per channel, if not more, again it depends on how many watts they can handle), which would be alot bulkier and louder then the H2O rig, and you'd have to worry about condensation and insulation.
The only reason you'd need a compressor would be to get something cold (ie below ambient). If you want to cool something down to somewhere above ambient (or very close to ambient) there are much more efficient ways of doing it. I don't know maybe you need you're amp below zero, but I truly doubt it for some reason :). |
|
|
| Kilowatt |
| Actually, I said antifreeze in case the amp will ever be outside when it's below freezing, but 40% methanol would also keep it from freezing, and nothing would grow in it. |
|
|
| PassFan |
| Have you thought of what this thing is going to weigh? 7200 watts of output + 4000 watts wasted heat is going to mean about 12000 watts of power supply. A 2000 watt tranny can weigh in around 30 lbs. How are you going to carry it. Designing it to be rackmountable would be ideal. Maybe refigure how you set up your cabinets to your amps to create a more weight efficient setup? Just a thought, your the one that has to carry these things. Coming from a former musician/personal roadie, it might be worth it at first but your going to get tired of lugging these things around. Make it easy on yourself right up front. Calculate the weight, and set it up with wheels in road cases with a not to exceed weight. You'll thank yourself for it later. |
|
|
| AudioFreak |
PassFan,
The amp is Class B or AB so will not have much in the way of quiescent dissipation........
Also the supply is transformerless and each 900W amp will have a seperate power supply ..... |
|
|
| blmn |
| I think GRollins has made something like a watercooled amp. |
|
|
| Kilowatt |
Actually, it will have one power supply for all 8 channels. It will have a 60A breaker as an on/off switch, 4 150A discrete rectifier diodes, at least 8 200V, 31,000uF filter caps (not much for a brute like this, I know, but it will have a high PSRR, and besides, those caps are huge and expensive), smaller electrolytics, film caps, two MOVs, 20W bleeder resistors (I'd hate to see those caps discharge into a screwdriver!! There wouldn't be a srewdriver left, I'd be seeing spots and have a ringing in my ears for a week!), and possibly two big inductors mounted in a compartment under the rest of the chassis.
I wonder how much tubing I'd need in the heat exchanger for this thing? |
|
|
| AudioFreak |
| thats not gonna be anywhere big enough for this amp kilowatt.. on transients this amp is gonna load down like crazy if it doesnt have heaps and heaps of low resistance capacitance in the power supply and because all channels will peak @ the same time, the peak current draw will be in excess of 100A |
|
|
| Jeremy_Wolf |
KiloWatt, I don't know if anyone else has asked this exactly, but what are you going to use this thing for?? Simulating earthquakes, producing 5 hz tones at 140 db? This is a serious amount of power. Of course you could just be doing it for fun which is a good enough answer for me. But if you really need that much power, wouldn't it be cheaper to buy a comercialaly made amplifier setup? I know that a crown ce2000 is rated at 975 watts/channel into 2 ohms and they only cost about 550. Get 4 of these and you have your power requirements, it will only cost 2200 dolars for all 8 channels, all of the protection circuitry is built in, all of the cooling is built in, and to top it all off, there is a warrenty and they only wieigh about 40lbs each.
Not that I want to put a damper on the DIY spirit. Sometimes it is cheaper and wiser to go comercial. It's like designing a robot that can clean dishes when there are 200 dollar dishwashers for sale in every department store. But if this is a just for fun project, then good luck and take plenty of pictures for us all to see. |
|
|
| AudioFreak |
| the problem with the above mentioned amp is that it produces 975W into 2ohms.... Kilowatt needs that kinda power into a 12ohm load. |
|
|
| Kilowatt |
Also, the 7200W amp by itself will not cost more than $1000 US to build, not including the outputs I'm likely to blow up, of course.
How much capacitance should the power supply for this have? |
|
|
| PassFan |
| Where will you find a 12000 watt transformer for 1000.00? |
|
|
| Kilowatt |
| I won't, the power supply won't have a transformer. If you read this whole thread, you'd know that. AudioFreak said that just a few posts back. |
|
|
| haldor |
He's going to use the transformer up on the Utility pole.
I posted an explaination earlier in this thread.
Phil |
|
|
| Kilowatt |
| Yup, that's right, now, about how much capacitance would be recommended for the power supply? I think I'd need more if the amp had a lower supply voltage, but those 200V 31000 uF high grade, super low ESR, high ripple current caps are 3 1/2" in diameter by 5 5/8" tall and cost $182.75 each! Those caps are going to take up more space and more money than anything else in this amp. How many do you think I should have? |
|
|
| haldor |
Hi Kilowatt,
This question is really dependent on the design of your amp. Some designs are more tolerant of ripple on the rails than others.
I have seen rules of thumb that recommend at least 10,000 uf per 100 watts of power and I have also see people use 50,000 uF on a 50 watt Class A amp. The problem with rules of thumb is that they usually don't scale very well when you change the magnitude of the problem.
I would think that 100,000 uF would be enough for a 2000 Watt amp that is not intended for Hi-fidelity, but I have never built one this big. The simple answer is build one amp and try different amounts of filtering to see how it performs.
Given how many amps you're planning on building you could buy enough to make a couple of amps and the buy the rest once you know for sure how many you are going to need.
I know you are dead set on doing this, but did you know you can buy a brand new QSC PLX-3402 amp that will deliver 1700 Watts per channel into a 2 ohm load or 3400 Watts bringed into a 4 ohm load for about $1200. I don't believe you are going to be able to buy the parts to make your amps for less than this.
In quantity you might even be able to get thyem for less than $1200.
Phil |
|
|
| Kilowatt |
Hopefully this amp will have a pretty high PSRR. I'll just get more caps as I need like you say.
Is there anywhere on the internet I can get giant caps like I'll need for less than what I mentioned in my above post? I would prefer new ones, I don't like the idea of getting used electrolytics. |
|
|
| AudioFreak |
| If you know someone in the photographic processing industry, a good source of high voltage caps is available from the used disposable flash cameras ... the caps are usually 330V 120uF so you'll need alot of them .... try the local 1hr processing shop and ask kindly if you could have the used flash cameras... some shops send them back for recycling and may be reluctant but many just throw them in the bin.... just be warned that the cap is usually charged to just under 300V when you get the camera so be careful when you pull the back off... |
|
|
| Kilowatt |
| I'd need thousands of those caps, it's not much of an option. I think it would be too much trouble to use anything less than about 10000uF each, but thanks anyway. |
|
|
| AudioFreak |
| yeah you probably would want about 5000 of those caps ... but then again, they are free :) |
|
|
| Kilowatt |
5000 of them? I refuse to wire 5000 little caps together :) , and besides that, they are not all identical, so I would think that some of them would be destroyed much like if you try to wire up several BJTs in parallel with no resistors.
That's about 300,000uF per side, or about 20 of the big cans. Do you think that's about the right capacitance for this supply assuming fairly good PSRR? That would cost about $3800, not really an option. I need to find that kind of thing for cheaper! Anywhere I can get big caps, at least 200V, like that? |
|
|
| AudioFreak |
See you will face a few problems ... because you wont have a transformer in the power supply, now only do you have to be concerned about how much noise is on the rails, you've also gotta consider how much noise you'll be shooting back up into the mains..... i think you could have some very irrate neighbours if their TV reception suddenly goes down the toilet @ the same time that your amp starts making its presence know/heard/felt...
you've also gotta consider how much the rails will collapse under load and the line resistance ... all these things will have an effect on how much capacitance will be required.
most of the PA amps i've dealt with have used switching supplies i think ... tho they are a pain and murder to design properly, they certainly fix alot of the problems :) in very high power non audiophile applications. |
|
|
| blmn |
Killowatt,
Considering AudioFreak's words, you might consider to maintain this power supply caps with some charge, to reduce these "cool" efects over the mains when you turn on your "little" amps.
About how much capacitance you will need, the most important thing is to know something about the nature of the load of the power supply (class A :D , B, D etc) and its PSRR characteritics.
There is an rounded solution in the book Integrated Electronics by Milmann/Halkias wich might be enough for your very first approach on the question:
Vripple=Icc/(2*f*C) , for continuous loads.
Regards |
|
|
| Kilowatt |
| So, to check my math, 50A/(120Hz*0.12F)=3.47, so the ripple voltage would only be about 3.5V @ 50A with 120000uF right?. That's only about .02% of my supply voltage. Is that right? it seems a little too good to be true. Ought to be fine if it is though. |
|
|
| AudioFreak |
ripple @ the output should be well under 5mV so 3.5V ripple on the rails will need an awful good PSRR
even 1mV @ the output of those bass drivers will cause about 75dB of noise @ 1M
and 3.5V is about 2% of the rail |
|
|
| blmn |
There is something wrong here, because 3.47V is near 2% ripple if you consider, for example 180V per rail. 2% is a reasonable ripple under full load, I think, but, remember, your load is not constant and you might consider your mains capacity too.
It seems to me you can use these values and change it at your tests if you wish. someone above says something about making one or two units first to perform a fine tuning of your design. I think it is a wise advice.
Regards |
|
|
| Jeremy_Wolf |
Won't there be a huge inrush of current when you turn on the amp? Wont the caps draw a monsterous amount of current without limiting their initial turn on charging because of being directly connected to the mains? Also, what about the fact that a standard U.S. outlet is only capable of supplying 120V at 15 amps? Is that enough to quench your creations thirst? Are you going to have multiple circuits run to where the amp will be living?
What is this frakenstein project going to be powering, a small town? And do you have schematics of the amp you want to build :) If you do I want to see them cause I like overkill:D |
|
|
| Kilowatt |
The amp will plug into a standard American range outlet, which can supply 240VCT, 50A. Yes, there will be a huge inrush of current at startup, and it wouldn't surprise me if it makes the neighbors' lights dim for a second or so, but I will have 600V, 150A discrete diodes in the power supply, so it ought to handle it.
It will power 8 Audiobahn ALUM12X subwoofers as part of a tri-amped system easily capable of 140+dB @1M.
As for the schematics, I'm still designing it, but the front end will look a lot like the Leach Superamp with 20 MOSFETs instead of 8 BJTs, and the rest will look a lot like one of Randy Sloan's high performance MOSFET amps. I am hoping for a THD of less than 0.1%, and the amp will be capable of driving almost any speaker system you can throw at it. Not to brag, but hopefully, I will end up with one of the best super high-performance, painfully powerful high end class AB,B MOSFET amps that can possibly be built by an ordinary audiophile. :D |
|
|
| AudioFreak |
| man i want a 50A outlet!!! for my XA based amp! |
|
|
| Kilowatt |
| Huh, I have two in my house, but I don't see how I can give you one.:) Almost all American households have at least one, we plug our ovens and arc welders into them. We use very similar 30A outlets for clothes dryers, surely you must have one too. What do you do when you need lots of power? |
|
|
| AudioFreak |
| we have 240V 10A outlets..... not CVT ... for higher power he have 15A outlets... speacial 20A distribution plugs and 3-phase when higher voltage is needed. higer current is generally only available for industrial installations. |
|
|
| Kilowatt |
| Interesting. Anyway, on a different subject, I am finding it a little confusing trying to modify the Randy Sloan type topology to have series-arranged output devices like the Leach Superamp. This is because the Leach design has a predriver stage (Q14 and Q15, here's a link http://users.ece.gatech.edu/~mleach/superamp/ ) that does not exist in Sloan's design. This stage seems rather critical in the double-barelled OPS design of the Superamp, but I'm rather inexperienced with such things. |
|
|
| subwo1 |
| It seems like the Leech Super Amp is practically Class G. It seems like you could add a lower level power supply at 2/3 the level of the highest level. Then you connect diodes to the lower transistors to allow them to withstand reverse bias as the upper ones take over. The trade-off as I understand is some extra distortion as the storage charge on the lower transistors' bases prevents them from switching seamlessly. But the Hitachi amp that was using the technique did not spec with very poor distortion figures. But after the modifications, I doubt the amp could be considered low TIM any more. |
|
|
| Jeremy_Wolf |
| KiloWatt, those audiobahn alum 12's don't need "that" much power to play very loud. I know this because I have 3 of them in a linkwitz transform setup, and as we all know linkwitz transforms need a lot of power. I was using a crown ce1000 to power my three wich was a 4 ohm load (12 ohm each, then all in parrallel). It was a good matchup. I could exceed xmax at 20hz. I still use the same 3 audiobahn alum 12's now, but I use a CE2000 i picked up dirt cheap. The alum 12's are a very effeicent sub. I have been able to hit over 130db in my house at around 50 hz. These subs get way loud. I am also no where near maxing out the amp either. Either way your setup is going to be sweet. |
|
|
| Kelly McDonald |
Don't bother Jeremy. Kilowatt has already made up his mind. Even though his subs will reach excursion limits long before they get to 7200Watts, never mind that running an amp of this size directly off a 240V line is an increadibly dangerously bad idea, even after person after person has recommended trying somthing a little less collosal for a first DIY amp project. He is unwilling to listen.
I'm just waiting for the post where he tells us about the $500 in transistors that just went up in smoke, or how his house just burnt down and the insurance company won't cover a thing, or how he's now picking up bits of speaker from his room. God hoping that he doesn't end up killing himself. |
|
|
| Kilowatt |
| I don't think the SuperAmp could be considered class G in any way, because it only utilizes 1 set of supply rails and always draws from them. It just has the outputs and drivers in series so they don't see as much voltage. I'm just wondering where I should put the emitters of Q24 and Q25 to use this series OPS setup with a topology like the one at http://www.aussieamplifiers.com/sym-sch.htm, or other MOSFET designs, instead of the Leach topology, because those MOSFET amps don't have a predriver stage, which is what Q24,25 are in series with. |
|
|
| blmn |
AudioFreak,
About the ripple at full load, I think you might consider the class of the amp and its current draw characteristic to define a good or bad ripple characteristic for the power supply. Based on Kilowatt informations about the circuit, if I remember well, .005V will be near 100dB below the signal (considering peak to peak excursion). At quiescent levels, Icc will be low and the Vripple will follow it (70mV for Icc=1A, for example, using the same equation).
I really don´t remember for now, but in my prototype (made near 20 years ago :D ), I used 10000uF per rail and I had no problems with hum or noise at the output, I remember THD+Noise was below .1% in any case.
regards |
|
|
| Kilowatt |
Kelly, I just lost $500 worth of OP transistors, and they took $1400 worth of subs with them, and my house burned down too and the insurance won't cover it. None of that's actually true, of course, since I won't even be finished designing this thing by the time I have my Leach amps done, but at least you heard the post you've been waiting for. Now would you please not try to make this thread go where my other one on this project did. This isn't going to be my first project, it's just that I've already started thinking about how I'm going to do it, and I know some people have said not to do it, but many have actually helped me on it too, rather than just telling me to drop it and go into quilting or something. You seem to think that having 300V wiring in a closed, grounded box with isolated hookups is more dangerous than motorcycle stunt driving. And I'll probably get more Audiobahns so they won't xmax so easily, but after the amp's done. Anyway, what do you expect me to do? Just forget about the whole thing? Let me tell you something... NO! Would you? Also, I believe Jeremy rather likes this idea, with quotes like "If you do I want to see them cause I like overkill:D " and "Either way your setup is going to be sweet."
I suggest that we all forget about Kelly's post because it seems to aimed at breaking down our discussion.
Anyway, it sounds like I should test the amps before I decide just how much capacitace I'll end up having to have.
About my series OPS, shown below is a small section of what seems a reasonable topology for my amp. I am assuming that most people on the forum recognise where in the amplifier circuit this is, and can tell where all the loose ends lead. I would like to eliminate the stage highlighted in red, as it is normally not found in MOSFET amps, but I can't seem to find a way to because I don't know where to put the transistor that goes in series with it, the one with the voltage divider at it's base. In fact, I'm not even sure if this would be the proper way to put MOSFETs in series. Maybe there's something better. Any ideas? |
|
|
| AudioFreak |
| quote: | Originally posted by blmn
AudioFreak,
About the ripple at full load, I think you might consider the class of the amp and its current draw characteristic to define a good or bad ripple characteristic for the power supply. Based on Kilowatt informations about the circuit, if I remember well, .005V will be near 100dB below the signal (considering peak to peak excursion). At quiescent levels, Icc will be low and the Vripple will follow it (70mV for Icc=1A, for example, using the same equation).
I really don´t remember for now, but in my prototype (made near 20 years ago :D ), I used 10000uF per rail and I had no problems with hum or noise at the output, I remember THD+Noise was below .1% in any case.
regards |
Yep i didnt think about that @ the time :) |
|
|
| Kilowatt |
| What good audio design and analysis software can I get for a reasonable price? I've heard about Electronics Workbench, but it's hundreds of dollars, the pro version is thousands of dollars. It would be nice if the design software was less expensive than the amp I'm going to design with it. |
|
|
| R. McAnally |
Kilowatt,
Honestly I don't think there are any cheaper packages out there that are worth your time, or that have enough pin or part capacity for what you are doing. At times, I have been able to break larger circuits into smaller chunks (i.e. individual stages) and simulate with Pspice... but this doesn't always give you the whole picture..
My suggestion is to find someone who has a copy of what you need =)... good luck though! Be prepared to spend hundreds of dollars otherwise.
| quote: | Originally posted by Kilowatt
What good audio design and analysis software can I get for a reasonable price? I've heard about Electronics Workbench, but it's hundreds of dollars, the pro version is thousands of dollars. It would be nice if the design software was less expensive than the amp I'm going to design with it. |
|
|
|
| Geoff |
Randy
It is clear what you are suggesting, but it is not in keeping with the spirit of this forum to advocate a breach of copyright.
Geoff |
|
|
| R. McAnally |
I wasn't implying that copyrights need to be broken... is it illegal to use someone's computer?
| quote: | Originally posted by Geoff
Randy
It is clear what you are suggesting, but it is not in keeping with the spirit of this forum to advocate a breach of copyright.
Geoff |
|
|
|
| Kelly McDonald |
I'm not suggesting you give up on the project, but that you take a realistic approach towards achieving it. Its not a trivial or cheap project you are attempting here, and as may people have noted there are many, many, many factors that you will have to take into account. Many things that you can pretty much safely ignore when building a small amp, become very significant when building a large one.
Like I said before its very much like wood working, if you're just building a bookshelf out of plywood then you really don't have to worry about things like warping or shrinkage. However you go and build a solid oak table, then suddenly there are a whole buch of other factors you need to consider. Its also like building models, you don't have to worry about much with a snap-tite model, but you try and build a vacuform or resin model and suddenly it gets a lot more complicated.
Try building a few smaller simpler amps first, then play around with modding them. The cost of doing this is going to be a whole lot less than loosing 7200 Watts worth of parts.
300V inside a sealed enclosure isn't the big problem, as long as the inputs are opto-isolated and the speakers and amplifiers are in the same enclosure then its not as big a safety issue under normal operating conditions. However, note I said under normal operating conditions, since this is a prototype design and from what you said essentially your first attempt, normal operating conditions cannot be assumed. Design flaws, component drift, SOA and injected noise will all become serious factors. With a 50W amp, $20 bucks in output transistors and 40V a failure isn't going to be catastrophic. However you're talking several hundred dollars in outputs and a full 240V 50 Amp line, if this thing blows its going to go out in a very big way.
Belive it or not but I'm trying to help you, help you save alot of money and frustration. You know the number one reason people abandon hobbies is because they take on a very abitous project for thier first attempt, when it invariably fails they get discouraged and abandon it as a whole.
Also expect to spend several hundred dollars in equipment to build and debug this amp, you are going to need (or borrow), multimeters, modelling software, an oscilloscope, a signal generator would be nice too, so would a variable transformer that way you can power the thing up slowly (but a 240V, 50Amp variable transformer is probably going cost an arm an a leg). Ebay is a good source of cheap used test equipment. Still this project is going to cost quite a bit when all is said and done
However it is your life to do with as you please, but you are asking for help and opinions. I'm just giving you my advice, personally I would start by modelling the enclosures you are planning on using, this way you can determine how much power you are really going to need (you'll probably be surprised at how little it is), then instead of building one collosal amplifier (or bridging together a whole bunch of amps), you build a seperate amp for each driver. This way a single failure isn't going to take everything down and you have additional flexibility. Once you're confident with your ability, then you can try bridging them together to try and increase output.
This will be my last post on the matter and I wish you luck |
|
|
| Nelson Pass |
Last time I looked, you could still download a
student/demo version of MicroCap from
Spectrum software. I use an older version of
MicroCap on the few occasions I run simulations. |
|
|
| Kilowatt |
I'm not really after a certain dB level, I'm after as much as I can reasonably get with 7200W for bass, and I want a 7200W high end amp that I built myself, so the 7200W is in, but I can't afford to finish the whole system very soon, so I just plan on working on a few of the 900W channels for now, but I don't see how I could get very far on that without good design software, a distortion analyzer, and an oscilloscope, I'll probably get Electronics Workbench. I don't know when I'll have the money, so I'll get back to the thread when I have my Leach amps done. I'll get back to it as soon as I can.
By the way, Kelly, if you have something to say some time, by all means, go ahead, but if you don't want to post on the matter any more, ok, but thanks.
R. McAnally, "is it illegal to use someone's computer?" There is almost no chance that anyone in my hometown or even within a fifty mile radius has Electronics Workbench anyway. :) |
|
|
| PassFan |
Kilowatt:
Sorry about the tranny post. I didn't see Audiofreaks post.
Haldor:
You must have your threads mixed up. There is no explanation about using the pole mounted transformer in this thread. |
|
|
| AudioFreak |
PassFan....
Kilowatt will effectively be using the distribution transformer up on the power pole. |
|
|
| PassFan |
Audiofreak:
Do you think he'll get permission from the power company first.:D |
|
|
| AudioFreak |
| he wont really need it ... all the power that enters the home/building uses the distribution transformer.... all he will do it plug the amp into the power point.... |
|
|
| subwo1 |
| I, like Nelson, like MicroCap. It takes very little effort to use. If I had enough money I would buy the full version. |
|
|
| djss |
I too am in the process of building a multi-kilowatt amp. Mine is simply an 8 channel Leach SuperAmp. I plan to use it in 2 modes, 7.1 DTS (or 5.1), or bridged into 4 channels. The 4 channel bridged mode should be near your 7200W amp (combined). I got a 48 inch stainless steel rack + stainless steel drawers mounted on rails. The drawers are something like 10 inches tall and 28 inches deep. The final unit will house all 8 channels, a computer, digital processors, a water cooling system (with 2 tranny coolers), and over 30 120mm ball bearing fans on temp sensing switches.
I got the rack and drawers (50x38x24 inches, or close to that) and all the fans plus a big reel of coax signal cable and tons of other connectors and stainless rails for a total of $24. The water pump I got from my girlfriend's mom, who works at a garden hobbyist shop.
If you have any questions about my amp, feel free to let me know. I have found a way to QUICKLY identify and change a busted output transistor or an entire channel. In the end, I should be able to replace a busted channel and have the system running again in less than 2 minutes. I'm trying to put all the knowledge that I learned from my last 3 amps into this one.
As for your problems concerning design and test: I interned with an electrical engineer last year, and I can always go to him with troubles or requests to use his digital scopes or test equipment. Also, I can get help from professors at school on many issues.
That's just an overview, I'd be glad to answer any questions about the project.
PS....I'm taking MANY MANY MANY pictures this time around.
=DJSS= |
|
|
| Kilowatt |
| Great, I'll let you know if I have any problems. For now, I just need to worry about getting the water pump for my amp working. (it's liquid cooled, remember) It's a water pump from a little Toyota 4 cylinder engine, which happens to be separate from the block, unlike most water pumps on engines. I am (hopefully) going to run it with a 1/2 hp 3450 rpm Kitchen-Aid dishwasher motor (it's very quiet). I'm not sure how this motor was originally started, but I've had it running in capacitor-run mode. This setup doesn't have nearly enough starting torque for my pump. |
|
|
| JasonL |
ok people this is confusing, your all arguing about power and bridging to get more power BUT WHY what are you going to do with 16oo watts and up is there some kinda pride in having that much power, what is the poing unless your entertaining thousands of people and need all that power there is no point also you'd have to be ritch to run it every day it would consume so much power to turn it on and keep it running..
why not just build a class a ( or tube amp) with about 100 watts pure power and a realy good pair or efficiant speakers. .. that is what i am doing.. but my sub will have lotsa power ehhe and you dont need class a for a sub at all.. it is kinda pontless to me that you gus are wangint to have all this power and to do what with it if you were to listen to it youll be deaf in like no time.....
correct me if i am wrong but if you want loud go get a pair of head phones and CRANK IT UP..
Jason
P.s is anthony still around i emailed him and he never replyed.. to me.. |
|
|
| mgoedeke |
Kilowatt, I mentioned the LAB horn a while ago, and this would get you serious output at low distortion. You could even still use your 900W modules to power each horn. If used indoors, you’ll “only” need 2 of these monsters per side and it will give you decent performance to below 30 Hz (read up at http://www.prosoundweb.com/communit...ard=3&id=143250). If you’re hell-bent on building a 7200W amplifier ( and I applaud your courage), the horn reaches Xmax at about 300W @ 30 Hz (where a stack of 6 develops around 150 dB in half space), but I figure these will easily soak up transients of 1000W cleanly because excursion drops off dramatically above 30 Hz. Thermal power handling should be around 1200W per horn.
Another benefit of the horn, is that it presents a benign load to the amplifier and might help with the current drain situation. I would bet, that very few direct radiator systems would come close to such a beast and you get less distortion and far less compression for the same output (and investment in $$$). Only problem might be to build/move the LAB horns, though it should still be ok.
Wish you well on your ways
Martin Goedeke |
|
|
| Kilowatt |
Hey guys, I haven't posted in awhile, I'm about 1500 miles from my house right now on vacation in Arkansas. I'll be back next week.
Jason, I am going to have efficient speakers, so it will have LOTS of SPL. You can go louder with 7200W than 100W if you use the most efficient speakers for both (duh). And all 7200W of it will hopefully have no more than .01%THD, a design with similar topology gets .0038% at like 120W into 4 or 8 (I forgot which) ohms. The reason why I'm doing it has been covered, it's not made to pump 150dB into your ears, it's made for distance. I have yet to start getting deeply into the design, but I will later this summer. Anthony never replied to me when I emailed him before either.
Martin, where can I get LAB horns? How much are they, can I affored them? Also remember I want 12 ohms/module. Will they even work for me? (I hope they do!:D )
Thanks:) |
|
|
| Kilowatt |
Having nothing better to do (I can't work on my Leach amp until I can get some small drill bits), I have come up with a possible topology for my 900W amp modules, which I will officially call the "Kilowatt Amp." I haven't designed any amps so far, so don't expect it to be the finished product yet. Any ideas are welcome.
The input stage is almost the same as Slone's "figure 11.14," but using a more stable current source for the diff amps. The OPS is configured the best I could figure using Prof. Leach's series design with a CF MOSFET stage. I haven't calculated any component values yet, but I'll work on that soon and post an updated schematic. Right now I just want to know what people think of the topology.
I can't get it all to fit in the 100,000 byte limit, so I'll have to break it up. I've always had a lot of trouble trying to get things to fit. Is there a better way? |
|
|
| Kilowatt |
4
Again, I apologize for the inconvenient method of posting my schematic, but I don't know a better way to do it. |
|
|
| AudioFreak |
Kilowatt,
I hope this helps.
Here's what it looks like in one piece. |
|
|
| AudioFreak |
paint shop pro 7
I glued them all back together, resized the image then saved as gif because it's a better format for schematics. 10 minutes work. |
|
|
| Kilowatt |
| Never mind, I see it's a GIF file. |
|
|
|
