Amplifier design - need help
Happy New Year!
I am loking to design my first 100 W power amp and I need your help. What should I look for in the design if:
1. I want the impedance to stay between 5.5 - 13. Specifically I do NOT want to go lower then 5.5 ohms.
2. I want to have a headroom of exact 4 db. (100W RMS, no more then 250W peaks). (3db is not good).
Re: Amplifier design - need help
Just to clarify:
Your 5.5 ohms, are you talking about the load (speakers) you want to drive? Are you SURE none of the speakers that will be connected will ever go below 5.5 ohms (if that is what you mean);
On the power, do you realise that 4dB is HUGE? Is that based on anything or is it just a nice number you like (if so there is nothing wrong with it, just curious)? If you want 250W peak you need a power supply probably close to +/- 100V, and the power transistors that can withstand a minimum of 200V Vce, and still have good speed and a few other things. In terms of parts and cost that 250W peak power anp is a very different beast than a 100WRMS amp.
Thank you Jan.
Indeed, this specific amplifier will be designed to drive 8 ohm speakers that theoreticaly can go down to a minimum of 5.5 ohms. This amplifier will NEVER drive speakers that go lower then 5.5
On the power - I realize that 4db is huge. But in the same time I learned that most of NAD amplifiers from 80's and early 90's were designed with a 6db dynamic headroom!!!
4db is a quite a lot but I need it.
Whic power transistors will be those hat can withstand a minimum of 200V Vce, and still have good speed? The Hitachi Mosfets will do it? If not, which other would you recommend? (I personally prefer FETs).
"In terms of parts and cost that 250W peak power anp is a very different beast than a 100WRMS amp."
I realize that - that is why I need all the help I can get.
1) They could maintain that headroom only for a short time, and
2) Most of these were class G designs (proton also had some that were NED derived with similar specs).
I forgot - the plan is to have the power supply separated from the amp. I mean in a different box...
I didn't knew NADs were in Class G.
I think for my amp a single-ended Class A or
Class A up to 20W, then going into Class AB.
2SJ201 / 2SK1530 (Toshiba MOSFET)
BUZ901DP / BUZ906DP (Magnatec MOSFET)
GT20D201 / GT20D101 (Toshiba IGBT)
MJL4281 / MJL4302 (ON-Semi BJT)
These are propably some of the best transistors for audio output stages with nice specs.
OK, well I think you cannot predict the lowest impedance of a speaker easily (the spec sheet is just a part of the marketing dept. these days) but it wouldn't make much difference for your amp anyway, 5.5 ohms or 5 or 4.5.
There is a basic issue with a high headroom. If you design your amp for 250W peak, it will also deliver longer time power at 250W if the signal and speakers require it. So to play it safe you either have to very carefully set the V-I limiting circuit or design the output stage in such a way that it can deliver 250W for longer periods, which most certainly means multiple parallel pairs. I think the way NAD did it was with VERY soft supplies. In other words, when the amp had to deliver high peak levels at a longer time, the supply would sag a lot, 30% or so, and that would automagically protect the amp. But that is quite a balancing act with the supply caps and transformer capacity, I don't think you would want to do it.
Maybe you should just built yourself a 250WRMS amp ;) .
There is a reason why I want 4db dynamic headroom - beacause this way will emulate more the sound of tube amplifiers (that I love). That is why I insist: 100W RMS / 250W peaks.
Besides that - a 4db dynamic headroom will sound unrestrained and I will have less chances of running out of power.
I listen most of the time to classical music - and I measured changes in output of up to 30db!!! That is why a 4db dynamic headroom is required. To be able to manage the HUGE peaks and transients.
Regarding the impedance: I measured my speakers impedance - they go down to a minimum of 6.3 ohms.
When I said 5.5 ohms that would cover the lowest theoretically impedance that this speakers might ever take.
Your quest is not as difficult or exotic as some would have you believe,
The speakers 5.5 to 13 ohm could be considered a nominal 8 ohm load for all intents and purposes.
To achieve 250W (rms short term max) into 8 ohms, output will be 63.25V peak so you could likely manage with a +/- 70V supply under no load, with a suitably efficient source follower output stage ( as used in my GB300D 300W/8ohm amp).
As continuous power is only 100W but you are after some 20W Class A the PS will need to hold up to only about +/- 42V with 8 ohm loading. This means you can use what many would describe as an undersize transformer of maybe only about 140VA (per channel). A transformer of this size would have some 25% loss under peak demand. Capacitor banks should be only large enough to hold ripple down to a volt or so, but of sufficient quality to handle peak ripple currents.
The amplifier proper does need to have extremely good Power Supply Rejection and a very stable bias reference. If I can example my GB300D again, it uses a zener filter of the main supply down to 15V with which to bias the 2 BJT current sources that set the bias for the whole amplifier. This is very stable irrespective of supply vagaries whether 30V or 70V. Then you will need a PCB layout that does not have this very bouncy supply in any proximity to sensitive (high Z low signal) circuitry which may suffer stray pickup and degrade performance.
Your supply wiring should be loomed and spirited away perpendicular to the PCB at output stage proximity, twisted for field cancellation.
The continuous transformer power should be less than 40W while heatsinks will need to deal with 160W continuous and maybe double that for say 40deg C rise so fan cooled large finned extrusions will be needed unless you reduce the Class A need.
Good luck. Should prove interesting and informative.
|All times are GMT. The time now is 12:27 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Copyright ©1999-2016 diyAudio