Second attempt at Building my own amp from scratch
Ive been reading this forum and have greatly enjoyed all the experience of the members here.
Here is my 2nd attempt this time I really want to learn.
Criteria is a 100W amp into 8Ohm, with acceptable performance,
I dont expect a great result as this is a first and very new to me.
I wanted to keep it simple. I drew have 2 circuits,
Ive made a toner transfer pcb of the first circuit and will build it as
The second circuit is just a thought, to be more flexible at various supply voltages.
Ive only spent about 2hrs on second circuit.
Pls inform of changes I should make.
Ive tried to do the best I can, but I need advice.
The first circuit looks ok at first glance. The second is better.
The only danger is the bias pot, make sure it is in the right direction (max resistance) when you first switch it on or it will blow up the amp.
Bias pot looks correct and can be replaced with a fixed resistor once the value is established. A few things... The CCS for the IPS might be a bit low, the input transistors would like at least 1mA each. Ft drops off if collector current is too low, take a peek at Ft or GBW compared to Ic in a datasheet. Placing a pot in series with R2 will adjust the DC output offset by adjusting the IPS bias current, allowing the mirror to balance the difference and still keep equal degeneration in the IPS. Ex. If the DC output is +1V, Ve of Q2 is 1.6V and Ve of Q1 is 0.6V so Q1 tries to conduct more than Q2. Increasing the IPS CCS current allows more base current to Q3, making the output track further negative, turning on more Q2 and establishing a balance where the proper amount of base current is allowed to flow to Q3 so as to create 0Vdc output. Also Q3 needs an emitter resistor, ~=R4. R9 could be larger, ~2mA or so. When the output is driven completely to the +rail as in clipping, Q12 will saturate. Adding a base resistor ~1K to Q12 will prevent this saturating situation from effecting the IPS CCS. Miller Cdom, C3 seems to be a bit large. Also a cap, ~47 - 100uf, across R8 will help prevent cross-conduction of the output transistors at high frequency testing. Some local decoupling right at the output transistors wouldn't be a waste either.
The second requires a 470Ω resistor in series with the base of Q12, I would also probably add 4.7Ω to the bases of all the outputs.
Thanks guys, Ill make some changes based on what you guys have said for far.
First, if you don't have the books by Douglas Self and Bob Cordell then buy them!
My comments on the first circuit.
R3 could be much lower resistance to reduce noise. 1K at most, preferably less.
RF1 and RF2 similarly could be lower. Cost is extra power dissipation so need to ensure adequate power capability.
No output inductor?
C1 and CF2 are electrolytic capacitors. Read Self on this for a reasonable presentation of when capacitors matter and when it is "fetishism".
Second circuit similar except R15 and R10 could be increased to reduce noise.
Samuel Groner has an excellent website with a must-read paper on this topic.
BTW. What did you use to do the circuit boards?
Again thanks to all that have posted here, Im trying to implement most of the improvements suggested.
They are pricey here and Ill invest in another as finaces allow.
So I just used the arbitrary 100pf.
If the value can be calculated easily Id love to know how.
If its to complex :D, Ill trial and error it, when the amp is built.
Any more info here would be great.
lol, Ive never been lucky enough to write a program for one med/large microcontroller project that didnt need its fair share of fault finding.
Ive etched my pcb of circuit 1 and will be drilling and soldering it during the week, It wont be great, but Im still excited to build it.
Ill supply it with my newly made DIY Bench Supply for initial testing.
hehe I think the Supply cost me more to build than if I bought one.
Ill only etch circuit 2 after a good couple of revisions.
Thanks for all your help.
A couple of changes, after some calculations, hope its more right now.
Choose about 500kHz for low pass filter on input, and main pole.
Nearly done with Circuit One.
I have no Data on the heatsink I'm using.
So I'll Take Heat readings and work backwards to get its thermal resistance,
so I'll know for future, my supplier sells them for a reasonable price.
Circuit One is just a test circuit, I'm not intending to use It for Home.
Just A Project to help me understand some aspects.
Once I'm happy with Circuit 2, I'll be interested to compare it to Circuit 1.
Its been great fun, but I'll not have time to test it till monday.
Ill first supply with 30V-0-30V DC from my Lab Supply with Ilimiter set low.
then fault find if nessesary :headbash:
Here is the result of build 1.
Pcb Board was covered with plastic spray to stop corrotion while Im working with it,
flux is also not removed, so underside looks abit dirty.
|All times are GMT. The time now is 03:30 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio