Yeah, I'm kind of leaning in that direction. It's also slightly more well-behaved that way with a symmetric clip IIRC.
The initial definition of "bog standard" was "things that I have lying around". I guess we need to expand it a bit if we're addressing a wider audience.
I don't touch any semiconductor that's not from Mouser or DigiKey (or Farnell or Elektrokit or whatever the reputable vendor is in various parts of the world). Not since I was in a pinch and needed a 7812 quickly and stupidly got one from Amazon. It lasted for three minutes and then went on to put the unregulated 24V rail into my circuit. Never again. I gladly wait a few days and pay shipping.
I’d like to see a “real” amplifier that can be made exclusively with EBay/Aliexpress parts and still work. But it’s expecting too much. Closest I had was an all-Radio-shack sourced amp that put out a whopping 15 per channel at 4 ohms.
I always use the transistor VAS limiter, because it protects the VAS in the event of VI limiter activation.
I always use the transistor VAS limiter, because it protects the VAS in the event of VI limiter activation.
Roger that. It's back in the circuit. I hacked the diode back into the built amplifier from the original design that I'm currently using as a daily driver. But the new and improved version will have the transistor.
Once I've nailed down the design and produced a PCB, it would be a fun exercise to build two exact copies, but one with components for Mouser and one from AliExpress and test them side by side. Might make a good YouTube video, if nothing else, to stick in the face of the "bUt aLiExPreSS iS jUsT aS gOoD"-crowd. I've come across a few of those guy. You know, they successfully built a LED blink from AliExpress components and their next project is a 500W amplifier from some dodgy site that they're going to source entirely from AliExpress. It's funny. Whenever I ask them how it went, I never hear back.
Probably an utterly dumb question, but how do you measure quiescent current without the source resistors?
I simply watch current levels on my bench power supply ammeter and substract estimated drivers/VAS/input stage current (e.g. 40mA).
Yeah, I guess you could just turn it down to zero (if the pot allows that), measure the current call that the idle supply current without the bias.
And obviously I know how an ammeter works. I was looking more for best practice. Which you just gave me. 👍🏼
And obviously I know how an ammeter works. I was looking more for best practice. Which you just gave me. 👍🏼
Not really. IRFP240/9240 at these modest voltages are just fine with zener clamps and it is simpler. When those come on, they also spike the VAS current, but not by as much.
Yes, but don't I need some very small reverse voltage zeners for that? I tried a 5.1V zener in series with a 1N4148 (for reverse current protection) and clamping the gate voltage at around 6V still gives me 20A of current through a 0.1ohm impedance (to simulate an "almost short"). That's a bonfire.
DC, yeah, it would be a bonfire if not otherwise long term limited. Under AC excitation it would be half that on average in each FET and a fuse can take care of fire control. It’s been known to work with these FETs as well as laterals. The whole advantage of mosfets in the first place was supposed to be the short term thermal tolerance.
Yeah, that makes sense. Next item on the agenda is whether to put the zeners before or after the gate resistors. I've seen some design where they put them before the gate resistors, which kind of surprises me. Common sense tells me that they would have to pull less current if I put them after the gate resistor. Yes, the driver is current limited (thanks to the contraption we discussed yesterday), but it still feels like a better practice to let the resistors take the brunt of the current limiting.
Here's what it looks like right now. (Minus the redundant diodes. I removed them, but my computer crashed before I saved it)
Here's what it looks like right now. (Minus the redundant diodes. I removed them, but my computer crashed before I saved it)
The gate resistors need to be right at the gate lead in order to perform their primary function -suppression of local oscillations. If the clamps can be put in without compromising that short path then they go after the resistor. If not, then they go on the other side of it. Make them 1/4 watt resistors and they become effective fuses for a sustained short. You are supposed to be able to hear when something is amiss and take steps to correct it. In pro use you would always use a VI limiter to allow it to run in current limit for a while and not just go out altogether, overheat, or die at the first sign of trouble.
Yeah, my current board has the gate resistor 4mm from the gates and I'll try to keep it that way. I'm pretty happy with they way this looks now and I'm going to start on the PCB layout so we can take it for a spin and check what it sounds like.
Hey BTW, @wg_ski, what's your take on output coils on MOSFET amplifiers? I've heard people say they're not really needed for MOSFETs. but I'm not sure if that's based on any science. Without it, I have a 3dB down bandwidth of >1MHz, but it goes down drastically with a 3uH coil, for obvious reasons. Now, the bandwidth is still 300kHz, so it's very arguable whether it makes any difference, but the coil also takes up valuable PCB real estate. What's your take?
Don’t put the coil on the PCB. Wrap a 10 ohm 5W MOX (or 2W CC, same form factor) with #18 magnet wire and mount it on the speaker terminal. It’s not always needed, so inclusion can be left up to the builder.