Hi everyone and thanks up front for any help or advice!
I've just got a chance to restart work on my TDA2030A miniamp and I THINK I have figured out the problem after a lengthy day of building/troubleshooting/knocking my head into the wall.
Here are the symptoms: I build the amp with one chip and fire it up; it works great. Then I attach the second channel's chip to the PS and HMMMMMMMM (from BOTH channels). And the chips heat up. Alot. I switch off the power and check everything over (looks good). Then try again (same result). ARRGHHHH!!!
Okay, disconnect the new chip and try again. Music. No hum. No heat. 🙂
Obviously a bad chip, right?
So, rebuild 2nd channel around a NEW chip. And try again. HMMMMMMMM (from BOTH channels). etc.
Disconnect the new chip and try again. Music. etc.
Here's the REALLY STUPID QUESTION: Do I need separate 100uF (or 2200uF, which is the size I used) PS caps FOR EACH CHANNEL (each chip) or can I share a common power supply? When building with valves, I ALWAYS use a common power supply (being too stingy to build discrete PS for each channel).
Please note that I DID wire .1uF caps directly to the appropriate pins of each chip!
If so, do I need different rectifier bridges for each chip? Different windings on the PS trafo? Or can I just share the PS as built and add some 100uF caps for decoupling adjacent to each chip?
Thanks again, and sorry for wasting your time with a really dumb set of questions. Frankly, if it's the way I think it is, I'll probably keep the PS module as is and go to a twin channel chip like the LM4766 for this one (the PS does have enough current to back up it's limited voltage swing for the 4766, even into a 4 ohm load).
Ja mata,
Morse
I've just got a chance to restart work on my TDA2030A miniamp and I THINK I have figured out the problem after a lengthy day of building/troubleshooting/knocking my head into the wall.
Here are the symptoms: I build the amp with one chip and fire it up; it works great. Then I attach the second channel's chip to the PS and HMMMMMMMM (from BOTH channels). And the chips heat up. Alot. I switch off the power and check everything over (looks good). Then try again (same result). ARRGHHHH!!!
Okay, disconnect the new chip and try again. Music. No hum. No heat. 🙂
Obviously a bad chip, right?
So, rebuild 2nd channel around a NEW chip. And try again. HMMMMMMMM (from BOTH channels). etc.
Disconnect the new chip and try again. Music. etc.
Here's the REALLY STUPID QUESTION: Do I need separate 100uF (or 2200uF, which is the size I used) PS caps FOR EACH CHANNEL (each chip) or can I share a common power supply? When building with valves, I ALWAYS use a common power supply (being too stingy to build discrete PS for each channel).
Please note that I DID wire .1uF caps directly to the appropriate pins of each chip!
If so, do I need different rectifier bridges for each chip? Different windings on the PS trafo? Or can I just share the PS as built and add some 100uF caps for decoupling adjacent to each chip?
Thanks again, and sorry for wasting your time with a really dumb set of questions. Frankly, if it's the way I think it is, I'll probably keep the PS module as is and go to a twin channel chip like the LM4766 for this one (the PS does have enough current to back up it's limited voltage swing for the 4766, even into a 4 ohm load).
Ja mata,
Morse
You've got 1 channel that works great and another that gives problems when connected to the power supply with the 1st channel, right? Try with just the channel that is giving you problems. It sounds like you either have too much ripple on the rails when using 2 chips or a ground loop or oscillation.
Thanks AudioFrank;
Yep, it looks like oscillation - more testing showed it to be some sort of motorboating (no o-scope here anymore or I would've spotted it earlier). That screams "power supply decoupling problems". 🙁
I think that next up on this project will be a rebuild using the LM4766, since I can then use the same chassis and PS board. Since the TDA2030A's are already "pin to pin" wired, I'll just toss 'em in my junk box and keep 'em for one of these days when I have some time on my hands and am thoroughly bored....
Now I just have to figure out how to integrate the 15 pin LM4766 with perfboard. Think I can (gently!) bend the pins to fit through the holes (after opening up the holes with a drill)? I'd like to build this using perfoboard to hold everything in place, yet continue using a 'pseudo point to point' scheme beneath the board. That's a style of construction I'm comfortable with from scratchbuilding valve gear. Any ideas?
Thanks again.
All the best,
Morse
Yep, it looks like oscillation - more testing showed it to be some sort of motorboating (no o-scope here anymore or I would've spotted it earlier). That screams "power supply decoupling problems". 🙁
I think that next up on this project will be a rebuild using the LM4766, since I can then use the same chassis and PS board. Since the TDA2030A's are already "pin to pin" wired, I'll just toss 'em in my junk box and keep 'em for one of these days when I have some time on my hands and am thoroughly bored....
Now I just have to figure out how to integrate the 15 pin LM4766 with perfboard. Think I can (gently!) bend the pins to fit through the holes (after opening up the holes with a drill)? I'd like to build this using perfoboard to hold everything in place, yet continue using a 'pseudo point to point' scheme beneath the board. That's a style of construction I'm comfortable with from scratchbuilding valve gear. Any ideas?
Thanks again.
All the best,
Morse
... how about only have a few pins go down through holes, and bend others to stay on the "top" side. that way you may get a spacing to line up with the perf better. (don't forget to try it on an angle.) the idea is just to keep the thing from dancing around, right? fix a few points and then be traditional on the rest. (... with a magnifying glass, apparently. 18mm wide 15pin dual? yikes.)
you might look at this link for more ideas. scroll to the bottom at the photos numbered 1 through 5. chap straightened his pins and then wrapped them around one edge of his perf board. maybe a combination of techniques could get you something manageable.
http://home.ca.inter.net/~cfraser/Gainbuilder.htm
you might look at this link for more ideas. scroll to the bottom at the photos numbered 1 through 5. chap straightened his pins and then wrapped them around one edge of his perf board. maybe a combination of techniques could get you something manageable.
http://home.ca.inter.net/~cfraser/Gainbuilder.htm
oh, and for bending the pins: you might try something like the nylon tube out of a cheap bic pen. use a chunk of that as a sleeve so you have something to put your fingers on and to control just where the bend point is.
i'd think you'd want to keep the sleeve short, too. last thing you want is leverage. better to keep it just long enough to hold, so you get maximum feel of how much pressure you're applying. do a little bend, then slide the sleeve out a bit and continue the bend farther up to keep from getting too sharp.
or something. experiment on some wire and chips you don't like first.
i'd think you'd want to keep the sleeve short, too. last thing you want is leverage. better to keep it just long enough to hold, so you get maximum feel of how much pressure you're applying. do a little bend, then slide the sleeve out a bit and continue the bend farther up to keep from getting too sharp.
or something. experiment on some wire and chips you don't like first.
Thanks ofb;
WOW! I can tell that I have a bit to learn about proper building technique for these solid state thingies (at least from AC mains - my battery powered OPA2134's have never given me problems!). The spacing between the caps and the chip is TINY in those pics! OTOH I've been making my zoebel even tinier (1/8 watt resistors and 50V ceramic discs are small enough to pack between pins 2 and 4, and a single 1/8 watt resistor for the NFB loop resistor - it's a good thing I'm nearsighted....).
Thanks also for the tips on pin bending; I've just been using a pair of needle nose pliars to hold the pins next to the chip body then carefully bending the pins with my fingers (straighten pin 4, then open up pins 3 and 5 enough to cram 1N4007's between 3 and 4 and between 5 and 4).
Given the small size of heat sink and cool operation with both chips on a common 5"x5" aluminium chassis, I may just try one last time with TDA2030A's and use a section of perfboard to hold the PS decoupling caps and a couple of the more awkward outboard components (there's a resistor/cap series off pin 2 of the TDA2030A that just screams "bend me/break me"). Then attach each TDA2030A to one of my smaller heat sinks (they're about 1" by 1/2" with some pretty good sized fins - I've finally got a pretty good supply of the bigger sizes of heat sink but there's no way they'd FIT in that small chassis....).
Now for the last thing I'm a little hazy on. Okay, I need 100uF decoupling caps as well as .1uF decoupling caps FOR EACH CHIP and they'd better be darn' close to the chips, too. BUT, is there any problem with keeping the 2200uF PS filters that are on my existing rectifier/PS board?
Thanks again and all the best!
Ja mata,
Morse
WOW! I can tell that I have a bit to learn about proper building technique for these solid state thingies (at least from AC mains - my battery powered OPA2134's have never given me problems!). The spacing between the caps and the chip is TINY in those pics! OTOH I've been making my zoebel even tinier (1/8 watt resistors and 50V ceramic discs are small enough to pack between pins 2 and 4, and a single 1/8 watt resistor for the NFB loop resistor - it's a good thing I'm nearsighted....).
Thanks also for the tips on pin bending; I've just been using a pair of needle nose pliars to hold the pins next to the chip body then carefully bending the pins with my fingers (straighten pin 4, then open up pins 3 and 5 enough to cram 1N4007's between 3 and 4 and between 5 and 4).
Given the small size of heat sink and cool operation with both chips on a common 5"x5" aluminium chassis, I may just try one last time with TDA2030A's and use a section of perfboard to hold the PS decoupling caps and a couple of the more awkward outboard components (there's a resistor/cap series off pin 2 of the TDA2030A that just screams "bend me/break me"). Then attach each TDA2030A to one of my smaller heat sinks (they're about 1" by 1/2" with some pretty good sized fins - I've finally got a pretty good supply of the bigger sizes of heat sink but there's no way they'd FIT in that small chassis....).
Now for the last thing I'm a little hazy on. Okay, I need 100uF decoupling caps as well as .1uF decoupling caps FOR EACH CHIP and they'd better be darn' close to the chips, too. BUT, is there any problem with keeping the 2200uF PS filters that are on my existing rectifier/PS board?
Thanks again and all the best!
Ja mata,
Morse
Yep, that problem has happened to me with the lm1875. I was running both channels off of a 1A +-20V supply with only 2000uf of power supply caps. One channel or the other would work, but if I tried both at the same time, they would work for low volumes, then one would start to buzz, then get louder and louder, and warmer and warmer, then they would give up and stop working until I reset the power.
-Mike
-Mike
Thanks SoundNerd;
>>>...with only 2000uf of power supply caps...<<<
Hmmmm, "only" 2000uF - and here I thought I was going overboard, when the original spec called for 100uF! I don't think I've ever used more than 200uF in a pi filter for one of my valve projects.....I guess this is what happens when a valve guy starts the learning curve on chips - we do the same things we are used to having work with valves, and have to discover that some of 'em don't work with chips. I wonder if it works both ways, and that's why valves have an evil repute with some solid state guys of my acquiant - there doubtless are things that work fine with chips but are the kiss of death to a valve based circuit!
>>>...One channel or the other would work, but if I tried both at the same time, they would work for low volumes, then one would start to buzz, then get louder and louder, and warmer and warmer, then they would give up and stop working until I reset the power...<<<
Interesting. With the TDA2030A, it would roar from first power on. That implies to me that the LM1875 is an inherently more stable chip.
After sleeping on it, I've decided to rehab the chips in my basic amp (by removing the original power leads and .1uF caps) then go to the "outboard circuit board" idea, using 100uF decoupling caps in parallel with some slightly better .1uF decoupling caps. I won't have a chance to rebuild it for a while though. Today is my last really "free" day for a while, but at least I can clean everything up and get it ready for when I do have time.
Thanks again and all the best!
Ja mata,
Morse
>>>...with only 2000uf of power supply caps...<<<
Hmmmm, "only" 2000uF - and here I thought I was going overboard, when the original spec called for 100uF! I don't think I've ever used more than 200uF in a pi filter for one of my valve projects.....I guess this is what happens when a valve guy starts the learning curve on chips - we do the same things we are used to having work with valves, and have to discover that some of 'em don't work with chips. I wonder if it works both ways, and that's why valves have an evil repute with some solid state guys of my acquiant - there doubtless are things that work fine with chips but are the kiss of death to a valve based circuit!
>>>...One channel or the other would work, but if I tried both at the same time, they would work for low volumes, then one would start to buzz, then get louder and louder, and warmer and warmer, then they would give up and stop working until I reset the power...<<<
Interesting. With the TDA2030A, it would roar from first power on. That implies to me that the LM1875 is an inherently more stable chip.
After sleeping on it, I've decided to rehab the chips in my basic amp (by removing the original power leads and .1uF caps) then go to the "outboard circuit board" idea, using 100uF decoupling caps in parallel with some slightly better .1uF decoupling caps. I won't have a chance to rebuild it for a while though. Today is my last really "free" day for a while, but at least I can clean everything up and get it ready for when I do have time.
Thanks again and all the best!
Ja mata,
Morse
Morse said:
Well, with valves you typically use much higher voltages than
in solid state, so you need less current for the same power.
That is for solid state your capacitors can be rated at a lower
voltage, but have to have a higher capacitance to be able to
supply more current. Roughly you could say that
C*Vmax should be about the same in the two cases.
This is an oversimplification of things of course, but I hope it
illustrates some of the reason why things don't just carry over
without adjustment.
Thanks Christer;
>>>...Roughly you could say that
C*Vmax should be about the same in the two cases...<<<
Hmmm, I'd never thought about it in quite those terms.
FWIW, I usually use either a 20uF->10H->100uF pi filter or a 100uF->R->100uF pi filter with valve projects (the latter being for my cheapies - chokes are pricey!), with a second stage pi filter for input valves in the event of more than 2 stages OR the use of RF valves (like 6922's). In any case, I'm maintaining about that same product of C*Vmax on my valve projects as I used here (these days I seldom go much over 350VDC on my project's B+). Frankly, it just strikes me that some of these chips are a bit unstable compared to the valves I'm used to, and my next gen chip designs will all take that into account.
Anyway, thanks again for the input. You've given me some food for thought. All the best!
Ja mata,
Morse
>>>...Roughly you could say that
C*Vmax should be about the same in the two cases...<<<
Hmmm, I'd never thought about it in quite those terms.
FWIW, I usually use either a 20uF->10H->100uF pi filter or a 100uF->R->100uF pi filter with valve projects (the latter being for my cheapies - chokes are pricey!), with a second stage pi filter for input valves in the event of more than 2 stages OR the use of RF valves (like 6922's). In any case, I'm maintaining about that same product of C*Vmax on my valve projects as I used here (these days I seldom go much over 350VDC on my project's B+). Frankly, it just strikes me that some of these chips are a bit unstable compared to the valves I'm used to, and my next gen chip designs will all take that into account.
Anyway, thanks again for the input. You've given me some food for thought. All the best!
Ja mata,
Morse
Morse,
just don't take that formula C*Vmax too literally. There are a
lot of things to consider, like how much ripple you allow,
what type of filtering you use etc. I just wanted to hint you why
component values don't carry over directly and that you do,
in a sense, trade voltage for current. It is a bit like comparing
the household mains to the car battery. In the first case you
have higher voltage and less current, in the other case you
have lower voltage and more current. (Then we typically have
AC in the mains and DC in the car, of course, but I am sure
you get the point).
just don't take that formula C*Vmax too literally. There are a
lot of things to consider, like how much ripple you allow,
what type of filtering you use etc. I just wanted to hint you why
component values don't carry over directly and that you do,
in a sense, trade voltage for current. It is a bit like comparing
the household mains to the car battery. In the first case you
have higher voltage and less current, in the other case you
have lower voltage and more current. (Then we typically have
AC in the mains and DC in the car, of course, but I am sure
you get the point).
Yes, chips do need more capacitance on the power suppply, otherwise ripple in the power can cause oscillation (at least i think thats what is happening). Mine acted weird in another way, also. If I ran just one channel, after a minute of use, the cone of the speaker had moved out about 1/2 inch, and when I turned it off and on again it started all over again.
I also designed it badly, since I mounted the chips on the heatsink, the circuitry on an IC board from radioshack, and used wires to connect the chips to the board. I did build it, though, back before I had a laser printer to make nice pcbs, and was too lazy to use a sharpie or those rub on transfers from radioshack.
I used 1,000uf per rail, which just kills sound quality. Another amp i have, using the LM3886 for a sub started to sound great, with less distortion and tighter bass when i boosted the power supply caps from 2,000uf total to 8,800uf total (4400uf per channel). You really want more like 16,000uf to 20,000uf (8,000-10,000uf per channel) for the best quality and if you want the chips to not oscillate. Correct me if i am wrong, but i would think a cleaner power supply would equal less work for the transistors which would equal less heat. the only thing, though, is that the 2030 is by no means a chip with amazing sound quality, so you shouldn't even bother with much over 10,000uf total.
I still am going to make a stereo lm1875 amp with a nice pcb. Ill let you all know how it came out.
back to the 2030, I tried these chips, and they were nothing but problems for me. I really wouldn't reccommend them at all. Go with something from National. There chips all work and sound great (at least the ones ive tried).
I also designed it badly, since I mounted the chips on the heatsink, the circuitry on an IC board from radioshack, and used wires to connect the chips to the board. I did build it, though, back before I had a laser printer to make nice pcbs, and was too lazy to use a sharpie or those rub on transfers from radioshack.
I used 1,000uf per rail, which just kills sound quality. Another amp i have, using the LM3886 for a sub started to sound great, with less distortion and tighter bass when i boosted the power supply caps from 2,000uf total to 8,800uf total (4400uf per channel). You really want more like 16,000uf to 20,000uf (8,000-10,000uf per channel) for the best quality and if you want the chips to not oscillate. Correct me if i am wrong, but i would think a cleaner power supply would equal less work for the transistors which would equal less heat. the only thing, though, is that the 2030 is by no means a chip with amazing sound quality, so you shouldn't even bother with much over 10,000uf total.
I still am going to make a stereo lm1875 amp with a nice pcb. Ill let you all know how it came out.
back to the 2030, I tried these chips, and they were nothing but problems for me. I really wouldn't reccommend them at all. Go with something from National. There chips all work and sound great (at least the ones ive tried).
Hi Mike;
>>>...I really wouldn't reccommend them at all. Go with something from National. There chips all work and sound great (at least the ones ive tried)...<<<
REALLY! Hmmm, maybe I'll "rethink my rethinking" and go use the LM4766 in this amp after all....no point in building a loser.
>>>...I also designed it badly, since I mounted the chips on the heatsink, the circuitry on an IC board from radioshack, and used wires to connect the chips to the board...<<<
Well, um, actually that's pretty close to the situation here. I'm trying to stick close to techniques I'm familiar with (like point to point free-wiring).
Thanks again for the input! I can tell that a redesign of this one is in order.
All the best!
Ja mata,
Morse
>>>...I really wouldn't reccommend them at all. Go with something from National. There chips all work and sound great (at least the ones ive tried)...<<<
REALLY! Hmmm, maybe I'll "rethink my rethinking" and go use the LM4766 in this amp after all....no point in building a loser.
>>>...I also designed it badly, since I mounted the chips on the heatsink, the circuitry on an IC board from radioshack, and used wires to connect the chips to the board...<<<
Well, um, actually that's pretty close to the situation here. I'm trying to stick close to techniques I'm familiar with (like point to point free-wiring).
Thanks again for the input! I can tell that a redesign of this one is in order.
All the best!
Ja mata,
Morse
Just so you know, the lm4766 is the only chip i have had problems with. I was running it in bridged mode, and somehow blew the chip. It did sound good, though.
I would suggest using 2 lm3886s.
Im not trying to steer you away from the 4766, just telling you my story. in fact, i plan to build a 2 channel amp with it sometime (hopefully soon). ill post the pcb i come up with here.
-Mike
I would suggest using 2 lm3886s.
Im not trying to steer you away from the 4766, just telling you my story. in fact, i plan to build a 2 channel amp with it sometime (hopefully soon). ill post the pcb i come up with here.
-Mike
reread your post,
maybe chip amps just aren't tolerent enough to be connected like that?
somebody else could maybe input on this issue?
-Mike
maybe chip amps just aren't tolerent enough to be connected like that?
somebody else could maybe input on this issue?
-Mike
Hi Mike;
>>>...maybe chip amps just aren't tolerent enough to be connected like that?...<<<
You're definitely getting my attention here! Hmmm, what I'm thinking about now is using perf board to hold everything together and building a new PS/chip board (pardon the pun) that puts it all on together, with the decoupling caps as close to the chip as possible. I'll probably try 26ga wire too, since the heavier stuff is proving too stiff for use with delicate chip pins. The LM4766 is still my first pick, just because I have some on hand. I'll have to be mega-careful about oscillation though - I blew up a TDA2030A the same way as you describe (good thing I was wearing glasses!).
I guess this is why it's called "a learning experience"!
Ja mata,
Morse
>>>...maybe chip amps just aren't tolerent enough to be connected like that?...<<<
You're definitely getting my attention here! Hmmm, what I'm thinking about now is using perf board to hold everything together and building a new PS/chip board (pardon the pun) that puts it all on together, with the decoupling caps as close to the chip as possible. I'll probably try 26ga wire too, since the heavier stuff is proving too stiff for use with delicate chip pins. The LM4766 is still my first pick, just because I have some on hand. I'll have to be mega-careful about oscillation though - I blew up a TDA2030A the same way as you describe (good thing I was wearing glasses!).
I guess this is why it's called "a learning experience"!
Ja mata,
Morse
Layout, Layout, Layout............
Do TDA2030 need to be insulated from the heatsinks ???.
If this is not the cure, grounding techniques will be your problem.
Speakers must return directly to the center of the power supply, and signal earths all need their own path to the psu central ground.
In high gain circuits, any little bit of earth reference modulation will affect/effect performance and stability.
This earth modulation can cause transient instability, or conversely a transient limiting/compression characteristic.
Fine juggling layouts between these two conditions gives 'voice' to an amplifier, and defines the dynamic behaviour.
Eric.
Do TDA2030 need to be insulated from the heatsinks ???.
If this is not the cure, grounding techniques will be your problem.
Speakers must return directly to the center of the power supply, and signal earths all need their own path to the psu central ground.
In high gain circuits, any little bit of earth reference modulation will affect/effect performance and stability.
This earth modulation can cause transient instability, or conversely a transient limiting/compression characteristic.
Fine juggling layouts between these two conditions gives 'voice' to an amplifier, and defines the dynamic behaviour.
Eric.
Reply
Sorry Folks !
Amp_man_1 has a "FAILED STATE OF MIND" in case of CHIP AMPS.
Therefore i cant help u anymore.
Sorry Folks !
Amp_man_1 has a "FAILED STATE OF MIND" in case of CHIP AMPS.
Therefore i cant help u anymore.
Reply
something about INTEGERATED CIRCUITS POWER AMPS
They are like some black creatures with somany legs and ofcourse there is a hidden phenomena occuring inside it which i dont know.
something about INTEGERATED CIRCUITS POWER AMPS
They are like some black creatures with somany legs and ofcourse there is a hidden phenomena occuring inside it which i dont know.
And can we say about power darlingtons; They are extra terrestrial black tripods in fact! So we cannot know a lot about them! May be they are here to distact our world!
- Status
- Not open for further replies.