Santiago Alvarez (electronics engineer, JVM, YJM, AFD...)

santiall said:
well, I guess the main roadblock are the users that only want very traditional/conservative designs :D
Click to expand...

Do you have anything unique you’ve built for yourself? I’m just curious how much ground there is still left to be uncovered.
 
headcrash said:
Hey Santiago!

I just bumped into the schematic of the JVM1 on my hard disc.
Its FX loop takes one unused triode stage as AC CF buffer for the send, return simply feeds the PI.
When implementing this in a "proper" amp (50W or 100W), of course you would need to take down the send level quite a bit, to feed stompboxes, but the gain of a typical PI should make up for this. I could even put a 1M pot after the return, before the PI, as global master (maybe linear instead of log). This would also prevent hiss/noise of the FX device being amplified by the gain of the PI, because it would be taken down by the global master first.
Also the input impedance of the PI should be high enough to compensate any long cable runs coming from the FX device output, which in turn usually is low impedance, so ...

Is there something I haven't taken into account yet...?


View attachment 9035
Click to expand...
yeah, yeah, a few comments:
- the FX send signal is attenuated before the buffer. You can see R55 (33k) being the lower part of the attenuator of the signal that comes either from the clean or OD channels. In the JVM1 the loop isn't really intrument level, a bit higher so I can avoid a recovery level boost and the signal is big enough to properly drive the power amp.
- The FX return impedance is actually 100k (roughly, R57). The PI input impedance is approx 2M (R57 being bootstrapped and that'll be in parallel with R57. Anyway, more or less 100k which is a nice value for any FX processor output. The cable loading and stuff is taken care from the low output impedance of the FX processor, the 100k here is good enough to not load the FX processor output and also not too big to keep noise under control.
- Although the phase splitter has some gain, that's 'swamped' by the power amp feedback. The total gain from the FX loop to the speaker output is defined by the feedback, roughly 8x or so here. It is taken from the 16Ω output, 1W@16ohm -> 4VRMS so 0.5VRMS on the FX return should be enough to fully drive the power amp. Since the JVM1 doesn't rely on power amp distortion there you go, good enough for pedals.

A 100W amp would be the same but definitely you'll need a return boost to drive the power amp properly.
hope it helps!
 
Whizzinby said:
Do you have anything unique you’ve built for yourself? I’m just curious how much ground there is still left to be uncovered.
Click to expand...
there is always a pile of unfinished-and-never-to-be-finished projects on the desk... lately I've been experimenting quite a lot with getting power amp distortion without having to use a huge resistive/reactive load and not using any sort of power scaling either. Just testing some ideas of a different way to design the power amp. All tube of course :geek:
 
santiall said:
there is always a pile of unfinished-and-never-to-be-finished projects on the desk... lately I've been experimenting quite a lot with getting power amp distortion without having to use a huge resistive/reactive load and not using any sort of power scaling either. Just testing some ideas of a different way to design the power amp. All tube of course :geek:
Click to expand...
Sounds like interesting tinkering you're doing there :D
Looking forward to your revelations :grin
 
santiall said:
yeah, yeah, a few comments:
- the FX send signal is attenuated before the buffer. You can see R55 (33k) being the lower part of the attenuator of the signal that comes either from the clean or OD channels. In the JVM1 the loop isn't really intrument level, a bit higher so I can avoid a recovery level boost and the signal is big enough to properly drive the power amp.
- The FX return impedance is actually 100k (roughly, R57). The PI input impedance is approx 2M (R57 being bootstrapped and that'll be in parallel with R57. Anyway, more or less 100k which is a nice value for any FX processor output. The cable loading and stuff is taken care from the low output impedance of the FX processor, the 100k here is good enough to not load the FX processor output and also not too big to keep noise under control.
- Although the phase splitter has some gain, that's 'swamped' by the power amp feedback. The total gain from the FX loop to the speaker output is defined by the feedback, roughly 8x or so here. It is taken from the 16Ω output, 1W@16ohm -> 4VRMS so 0.5VRMS on the FX return should be enough to fully drive the power amp. Since the JVM1 doesn't rely on power amp distortion there you go, good enough for pedals.

A 100W amp would be the same but definitely you'll need a return boost to drive the power amp properly.
hope it helps!
Click to expand...
Thanks a lot Santiago for these helpful tips!

Urgently need to do my homework how to calculate/estimate the gain of a power stage with nfb. Of course the rather high gain of a typical PI is reduced quite drastically by nfb.
Since most modern amps don't rely on power stage distortion, the resulting gain might still be enough to amplify the few volts coming out of a modern-ish FX stomp box, like a Boss DD-7 or the tc toneprint stompboxes, or even something like a H9. These devices should be able to handle up to 5...6 Volt p-p signal swing, shouldn't they?
So considering this, even a 100 Watter could roughly deliver its rated power (5Vrms feeding the power stage would result in roughly 40Vrms, considering a PS gain of 8). However, people would be irritated, and consider such an amp as being not really loud, because they would be able to turn up the master volume real wide, use its full travel, finally :grin
It would be even "easier" to drive the PS, when nfb is not as high, using 8 or 4 Ohm outputs.
 
headcrash said:
Thanks a lot Santiago for these helpful tips!

Urgently need to do my homework how to calculate/estimate the gain of a power stage with nfb. Of course the rather high gain of a typical PI is reduced quite drastically by nfb.
Since most modern amps don't rely on power stage distortion, the resulting gain might still be enough to amplify the few volts coming out of a modern-ish FX stomp box, like a Boss DD-7 or the tc toneprint stompboxes, or even something like a H9. These devices should be able to handle up to 5...6 Volt p-p signal swing, shouldn't they?
So considering this, even a 100 Watter could roughly deliver its rated power (5Vrms feeding the power stage would result in roughly 40Vrms, considering a PS gain of 8). However, people would be irritated, and consider such an amp as being not really loud, because they would be able to turn up the master volume real wide, use its full travel, finally :grin
It would be even "easier" to drive the PS, when nfb is not as high, using 8 or 4 Ohm outputs.
Click to expand...
actually, you can pretty much estimate the power amp gain by dividing the feedback resistor by the 'lower tail' resistor. For example typical Marshall 100k/4k7: around 20 times. In real life isn't that much for to have an idea it is good enough. That's the gain to the output impedance tap that you chose, so if that was 4Ω, 100W@4Ω->20V, you only need 1V to fully drive the power amp into full power.

As a rough idea, usually stompboxed run on 9V power, that gives you an absolutely maximum pk-pk voltage of 9V in an ideal world. In realy life, probably we'd be talking around 6Vpk-pk but if for example it is something like an older distortion, that'll be limited to 1.2pk-pk... If you plane to use stomps I'd suggest to assume a max voltage of 1Vpk-pk, that's why you'll usually need recovery gain after the FX loop.

Also, bear in mind, that to fully overdrive something like a plexi, the power amp input will have a rather high voltage there, maybe 30V pk or more
 
Thanks again Santiago!!
So estimating gain of a push-pull power stage with nfb is basically the same like for a non-inverting opamp amplifier stage :facepalm Which the power stage in fact is, kinda.

In my assumptions I was thinking of modern hi gain amps, whose amount of gain is "readymade" before the power stage. Also I wouldn't think, one would seriously put an old tube screamer in a FX loop.
Normally modern stomp boxes even should handle conservative line levels, for example 6...8 dBu. So that often used level switch in FX loops should even be fine in the +4dBu position with many modern stomp boxes. Although I have to admit, it depends of course on the voltage level before the FX loop. So mostly it's better to have a send level and a recovery level pot.
 
headcrash said:
Thanks again Santiago!!
So estimating gain of a push-pull power stage with nfb is basically the same like for a non-inverting opamp amplifier stage :facepalm Which the power stage in fact is, kinda.

In my assumptions I was thinking of modern hi gain amps, whose amount of gain is "readymade" before the power stage. Also I wouldn't think, one would seriously put an old tube screamer in a FX loop.
Normally modern stomp boxes even should handle conservative line levels, for example 6...8 dBu. So that often used level switch in FX loops should even be fine in the +4dBu position with many modern stomp boxes. Although I have to admit, it depends of course on the voltage level before the FX loop. So mostly it's better to have a send level and a recovery level pot.
Click to expand...
that's what actually it is, a not-so-good op amp. Differential input with high input impedance, low impedance output. It is far from ideal but to guess things you can approximate it to an ideal one.
Yes, agreed it is better to keep the levels under control, not relaying on whatever the users plug in there, and then having the means to compensate in the return.
 
Hi Santiago what's the real differences between the two loops on the JVM? Is there any reason to use one over the other with a line level mutli fx?
 
SillyOctpuss said:
Hi Santiago what's the real differences between the two loops on the JVM? Is there any reason to use one over the other with a line level mutli fx?
Click to expand...
the original idea is that the series/parallel loop is 'the loop'. It is switchable, has level/mix controls and so on. The other one is more intended to allow connecting amplifiers together, or to use just the power amp part. That's why is is labeled PA INSERT / SERIAL LOOP. Of course you can use it also as a line-level only serial loop and connect rack stuff in there, the use the programmable loop as an additional level control
 
santiall said:
the original idea is that the series/parallel loop is 'the loop'. It is switchable, has level/mix controls and so on. The other one is more intended to allow connecting amplifiers together, or to use just the power amp part. That's why is is labeled PA INSERT / SERIAL LOOP. Of course you can use it also as a line-level only serial loop and connect rack stuff in there, the use the programmable loop as an additional level control
Click to expand...

Cool thanks Santiago that makes a lot of sense.
 
I'm not sure why it is, but I gotta say, I decided to try my pedalboard in the Insert/serial loop, and it works immensely better for me than the "normal" loop :idk my FX sound way better in there and I don't have any of the tone suck I was experiencing with the other
 
spawnofthesith said:
I'm not sure why it is, but I gotta say, I decided to try my pedalboard in the Insert/serial loop, and it works immensely better for me than the "normal" loop :idk my FX sound way better in there and I don't have any of the tone suck I was experiencing with the other
Click to expand...

I'm trying to work out the best configuration for my board at the minute. I'm having some issues tbh.
 
SillyOctpuss said:
I'm trying to work out the best configuration for my board at the minute. I'm having some issues tbh.
Click to expand...

I'd say give both configs a go just in case.. I went almost a year before trying the insert loop and the whole time the closest thing I had to a complaint with the amp was the loop. Seems a bit counterintuitive especially with what @santiall said above but I'm fully satisfied now :idk
 
spawnofthesith said:
I'd say give both configs a go just in case.. I went almost a year before trying the insert loop and the whole time the closest thing I had to a complaint with the amp was the loop. Seems a bit counterintuitive especially with what @santiall said above but I'm fully satisfied now :idk
Click to expand...

I've been trying both tbh and was happy at home volumes but have rehearsed twice now with it at gig volume and it's just not working for me. I think I need to try different input and output levels on my stomp.
 
spawnofthesith said:
I'm not sure why it is, but I gotta say, I decided to try my pedalboard in the Insert/serial loop, and it works immensely better for me than the "normal" loop :idk my FX sound way better in there and I don't have any of the tone suck I was experiencing with the other
Click to expand...
Sure, nothing wrong, you just lose the flexibility of the other loop (ability to switch it in/out) and the levels can be tricky as it is easy to overdrive many effect pedals in the serial loop.
Out of curiosity, what do you have connected and how did you use the programmable loop?
My experience is that 99% of the issues I have been asked since the amp was launched are due to an incorrect/ non-optimal setting of the MIX potentiometer.
Many people think you will control the amount of effect with the MIX on the back, but that's just not how a parallel loop works.
Having said so, I changed the loop in the HJS precisely because it is obvious that it is not a friendly loop in the sense that it is difficult to set in an optimal way.
 
santiall said:
Sure, nothing wrong, you just lose the flexibility of the other loop (ability to switch it in/out) and the levels can be tricky as it is easy to overdrive many effect pedals in the serial loop.
Out of curiosity, what do you have connected and how did you use the programmable loop?
My experience is that 99% of the issues I have been asked since the amp was launched are due to an incorrect/ non-optimal setting of the MIX potentiometer.
Many people think you will control the amount of effect with the MIX on the back, but that's just not how a parallel loop works.
Having said so, I changed the loop in the HJS precisely because it is obvious that it is not a friendly loop in the sense that it is difficult to set in an optimal way.
Click to expand...

Loop switching not so much a problem for me as I tend to never do that haha. I have not run into any issues overdriving anything in there

The bottom row 3 pedals here are all I have running in the loop
4-B6-D1-FB6-D7-DD-40-B2-BB76-A788-A17667-C7.jpg
,

I experimented pretty heavily with the mix knob, it seemed like two low and my delay sound I wanted was compromised and too high the tone of the amp was compromised. I kind of like the lack of a knob in the other loop as it seems to just work for me lol
 
You must log in or register to reply here.
Back
Top