the rail fuses located after the smoothing bank can be surprisingly low value.
What is the maximum rms current into the nominal speaker load?
for a 100W into 4ohm amplifier, Irmsmax = sqrt(max power / Rload) = sqrt(100/4) = 5Arms.
The fuse can be F2.5A and the amp can be full power tested without the fuse blowing.
This is based on Quasi's recommendation of half of the rms current.
I prefer half of the peak current, but I have checked a couple of amplifiers to Quasi's and the fuses do not blow.
An F2.5A before the main smoothing bank will blow repeatedly at start up.
It is worth checking that the output offset current (not voltage) is near zero when either of the rail fuses are removed.
Indianajoe, I'm running Linux as well and the simulation is LTspice for windows running on Linux with WINE.
The main reason manufacturers stopped using single-ended-supplies was to eliminate the cost of the expensive coupling capacitor. There's nothing stopping you from using caps to decouple the DC from your speakers if that's the protection you want. You'll need to get a couple of electrolytics and connect them back-to-back to make the cap non-polar. Then connect that series capacitor network in series with the speakers . Each capacitor will have to be rated for the maximum voltage swing. So you could get a couple of these Rubycon 10k uF 100V caps at $20 each and have yourself a quick and easy speaker protection circuit:
RUBYCON|100LSQ10000MRI36X118|CAPACITOR ALUM ELECT 10000UF, | us.element-14.com
Quick and easy, but not cheap. You'll need two sets of these (4 caps total) for a stereo amp. That's eighty bucks.
In the circuit I posted, Q3 is the drive transistor. Q4 is the conventional VI limiter set to a very low current on the plus rail. Q5 gates Q4 to disable it when Vout is large indicating a normal speaker load. Q6 & Q7 are mirrors of Q4/5 for the negative rail.
The main reason manufacturers stopped using single-ended-supplies was to eliminate the cost of the expensive coupling capacitor. There's nothing stopping you from using caps to decouple the DC from your speakers if that's the protection you want. You'll need to get a couple of electrolytics and connect them back-to-back to make the cap non-polar. Then connect that series capacitor network in series with the speakers . Each capacitor will have to be rated for the maximum voltage swing. So you could get a couple of these Rubycon 10k uF 100V caps at $20 each and have yourself a quick and easy speaker protection circuit:
RUBYCON|100LSQ10000MRI36X118|CAPACITOR ALUM ELECT 10000UF, | us.element-14.com
Quick and easy, but not cheap. You'll need two sets of these (4 caps total) for a stereo amp. That's eighty bucks.
In the circuit I posted, Q3 is the drive transistor. Q4 is the conventional VI limiter set to a very low current on the plus rail. Q5 gates Q4 to disable it when Vout is large indicating a normal speaker load. Q6 & Q7 are mirrors of Q4/5 for the negative rail.
Last edited:
The PV replacement basket includes the VC, cone, etc. They're surprisingly cheap. Peavey Scorpion SP10825RB Replacement Basket 0013870
I built two 6600 uf 100v back to back electrolytic capacitors and put them on the CS800s which is direct coupled split supply. It sounds funny on top octave piano at my 2V listening level. See http://www.diyaudio.com/forums/soli...ghlight=output+capacitor+subjective+objective thread. Zero crossing is chemical = slow. The djoffe mod ST120 with 3300 uf at ~35 volts DC output cap does not sound funny. Building 3 channels of that with better heat sinks. Glad to see AndrewT is still in, he is way better on circuitry than I am. But the history of what has happened is what I am getting at with capacitors, limit + 5% current long time is also bad. Maybe could combine your circuit with resettable PTC fuse of correct value for emmiter resistor; they do trip after a long 5% overload.
Guitar bands lose woofers, I've lost some $500 in unmatchable tweeters. Don't know why, ST70 has no reputation as being prone to oscillate.
Been figuring a way to package/stuff the individual channel fuses in the original ST120. Was going to put them before the rail capacitor as I thought others have done from the pictures, thanks for the tip AndrewT. Needs about a 2 amp MDL fuse for each channel after the rail cap.
Guitar bands lose woofers, I've lost some $500 in unmatchable tweeters. Don't know why, ST70 has no reputation as being prone to oscillate.
Been figuring a way to package/stuff the individual channel fuses in the original ST120. Was going to put them before the rail capacitor as I thought others have done from the pictures, thanks for the tip AndrewT. Needs about a 2 amp MDL fuse for each channel after the rail cap.
Last edited:
Gated VI Limiter Update
Working with LTspice, I was not able to make my gated VI limiter concept work properly in simulation. It introduced a lot of distortion, and I couldn't get it to significantly lower the current limit as the load increased. So I flipped the concept around and instead of cutting the amount of limiting, I boosted it. The implementation seems to work well in simulation. I was planning to add this limiting to an amp that was the subject of another thread, so I posted the schematics and my results there:
http://www.diyaudio.com/forums/solid-state/103632-cheap-100-150-watt-amp-14.html#post2471131
Thanks for your input.
Working with LTspice, I was not able to make my gated VI limiter concept work properly in simulation. It introduced a lot of distortion, and I couldn't get it to significantly lower the current limit as the load increased. So I flipped the concept around and instead of cutting the amount of limiting, I boosted it. The implementation seems to work well in simulation. I was planning to add this limiting to an amp that was the subject of another thread, so I posted the schematics and my results there:
http://www.diyaudio.com/forums/solid-state/103632-cheap-100-150-watt-amp-14.html#post2471131
Thanks for your input.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.