SE EL34 Plexi-verb

[Phil_S]

Lou, see page 3 of the thread. This is one before the final but good enough.

[Theashe]

Reducing R234 from 4.7k to 2k should give a bump a little over 6dB. Again, this is assuming you're not already close to your rails. You should be able to get around 12 Vpp out of a TL072 with 15V and ground for supply rails. If you have more headroom and the reverb is still quiet, go ahead and reduce that resistor even more. The ratio of R35/R234 sets the gain of the op-amp.

Alternatively, the output impedance of the BTDR-3 is 220 ohms and you've got a 1k resistor in parallel, so there's a small bit of attenuation there. You could get a small bump in dB (1.5dB from my simulation) by increasing R33 from 1k to 10k.

Lastly, raising C32 from 0.1uF to 0.15uF will buy a few dB in the midrange, but that may not be what you want.

I'd try playing with R234 first, since you can just clamp a 3.3k resistor across it to get around 2k.

[JazzGuitarGimp]

I would try the following:

Ditch C31 and P7. Leave pins 1 & 2 of the opamp tied together, and tie the output of the opamp directly to the input of the brick.

Remove R33 & C32. Break the two outputs of the brick apart, and connect each output of the brick, through its own 4.7K resistor to pin 6 of the opamp.

Replace R35 with a 100k pot for now - this will allow you to adjust the gain of the recovery opamp stage on the fly.

I'm still not sure it's a good idea to parallel the two sets of decay pot terminals into one pot. This may be causing an issue. It would be better, at least for now, to use separate pots and once the reverb is working, you can try using one pot again to see if it works acceptably.

Cheers,
Lou

[xtian]

You cannot use a single pot for both depth controls. I tried it. Awesome screaming feedback happened. Feature!

[JazzGuitarGimp]

Something else I would do: Since the brick can only accept a maximum input signal of 3Vp-p, and you've got the potential of feeding it 175Vp-p from the CF minus the tone stack loss, I would add a resistor, something between 10K and 20K from the point marked "Dry Signal In' to ground - that way you stand less chances of hitting the opamp too hard and damaging it.

[Phil_S]

I know the 5/15V power supply is good. That's on a separate board and doing what it should.

Naturally, I didn't build it exactly as drawn. These are all good ideas. It is now a real PITA to pull the board, but I will because it has to be modified. It could be several days until I get to this. Work is interfering. I'll try to post an "as built" schematic.

Thanks for being a great support team!

[John_P_WI]

Phil, I'm not second guessing, may i ask if you have verified your 15 and 5v rails under load at the module? As Xtian verified the module is quite the power hog. Voltage multipliers can be finniky under load.

[xtian]

As Xtian verified the module is quite the power hog. Voltage multipliers can be finniky under load.

I quickly posted a follow up--my multimeter battery was dying. After a fresh battery, I measured 40-60mA consumption by the BTDR-3H circuit. Not bad at all.

[John_P_WI]

Yeah i still think 60 ma for a ss device is a lot... even if it is only 0.3 watt....

[Phil_S]

Apologies are in order for misleading the technical support team. I know people volunteered their time to look and it was for naught.

I adopted Lou's BDTR-3 schematic almost exactly as he provided it, so many of the suggestions were not really in play any more. I am posting an updated full schematic, including updated voltages in the amp after modifying the B+ ladder.

My self-inflicted deviation from Lou's BDTR-3 schematic is the addition of 20KB trimpots on the outputs, connected in series to 22K resistors. Belton shows a 10K pot + 39K fixed resistor where Lou specified one 27K resistor alone on each output. I reasoned that the 22K to 42K range would allow me to dial in the output load on the brick and couldn't hurt. I didn't have 27K resistors on hand and figured this freelance "improvement" wouldn't hurt. I am now wondering if I fubared this mod when I wired it.

John, certainly your suggestion was worth checking. I verified the output, with the module plugged in, at 4.90V and 15.10V from the power supply board. There is no drop on the 15.10V, but the 5V supply drops from 4.98V to 4.90V. While I lack the experience to know if that's a problem, I allowed myself the latitude to guess that 4.90V is still OK. The Belton spec sheet specifies 4.5V min 5.5V max. I've also used the meter's continuity beep feature to verify the wiring to the connection point(s) on the BDTR-3 board.

While I'm looking at the BDTR-3 spec sheet, it shows supply current 60mA (typical) and 100mA (maximum). I have reasoned that using a 5V 3A transformer supply to the 5/15V board gives me virtually unlimited current for whatever is needed. (The transformer output meters at 6V unloaded and I am guessing the load is insignificant with respect to the capacity of the transformer, making the actual supply 6VAC.) However, being unfamiliar with 7805 and 7815 voltage regulators, I don't know if they are somehow limiting current in the output. Some assurance this isn't the problem would be appreciated.

My gut tells me there is a miswire or misconfiguration on the BDTR-3 board. I will be pulling the board and going over it yet one more time. This is my first experience in the solid state arena and I admit to being easily confused by the tiny work space and the layout challenges of a preconfigured PC board. I will be mapping the layout as part of the review, as that is the only real way to get proof of an error, so this could take me a while. I'm slow, see above comment about this being new to me.

Lou: I've got a question about the 1M input pot on your schematic. What is compelling you to spec such a large value pot? I'm sure my elementary understanding on this item is off the mark, but I'm thinking that it is too easy to shunt too much of the input signal to ground with a high value pot. Would a 1K pot function better, worse, or the same in this spot? Also, on the VGND, I should be seeing 7.5V?

I am reasonably sure an experienced person would have a more simplified layout or would have just made a custom PCB. I am really appreciating the wide open space you get with a turret board ;-} and have found a profound new respect for the people who do PCB work.

[JazzGuitarGimp]

Hi Phil,

Your assertions about the 5V winding's current capacity vs. the circuit's needs are correct. And 4.9V is within spec, though if you encounter a low line condition on your wall supply, you may fall out of spec. It would have been better to use a LDO regultor for the 5V supply - and you may want to think about replacing that in the future. But let's get the thing working first.

The 1M input pot is necessary because the source feeding the pot (the tone stack) is a high impedance source, so the load (input level trimpot) needs to be high impedance also. While playing, slowly adjust the pot until distooetion is heard in the reverb return, then back it off slightly. This should be done with the guitar's volume all the way up, and the amp's preamp controls as high as you expect to use them.

The virtual ground should be +7.5V.

I do see an error in the schematic, but it could be just that - a schematic error and not a wiring error. Pin 6 of the opamp should be tied to virtul ground, not ground.

Cheers,
Lou

[JazzGuitarGimp]

From the looks of it, you have a spare 12AX7 triode, yes? If the reverb circuit is wired correctly, you may need to add a triode gain stage after the reverb circuit. But I am pretty sure this should not be needed. The PI appears to have some gain, and all you need to drive the power tubes to full power is 28V p-p. The opamp will give you 12 V p-p, so just a bit of gain will get you up to 28V - and I'm pretty sure the PI has enough gain to get you there.

[Phil_S]

Lou: You are very kind not to hold it against me that I sent the crew on a wild goose chase with the out of date schematic. I am feeling bad about it.

There is no spare triode. There are 2 6EU7's (once again forgot to show this; 12AX7 is the same, different pinout) at the front end, with the 4th triode cathode feeding the tone stack. This is a SE amp, so there's no PI. I assume you are looking at the 6AQ7, a single triode, that functions as the recovery stage after the cathode follower t/s. The reverb send/return is between the t/s and the recovery triode.

I am just guessing without any real basis the 6AQ7 has a gain that is 25-ish. The power tube gets plenty saturated, so much so that I had to back off the guitar volume to about 5. The guitar has a pair of P90’s.

On the op amp, the symbol shows pin 6 on top and pin 5 on the bottom. I think (and will need to verify) I’ve got the output from the brick needs to go into pin 6. Pin 5 gets 7.5V from VGND. If I have this backwards, I am guessing it would explain what doesn't work? I chose the open triangle symbol for VGND as I didn’t have anything else that seemed appropriate. Look for the VGND label. Ground is 4 lines tapered into a triangle shape.

I should check VGND to verify 7.5V.

You mention using an LDO. The power supply board takes the 5VAC from the transformer and sends it through a tripler. The 7815 is fed 21V and the 7805 is fed 7V (tapped after one diode in the tripler.) It's regulated, so I don't expect line voltage will be a problem. 21V and 7V are metered output, not calculated.

Do you think the signal return from the reverb unit is somehow being swamped by the ordinary dry signal that goes across R14 (220K)? I had thought the 100K level pot deals with this. Well, one thing at a time. Let's deal with the op amp on the output side of the brick.

[JazzGuitarGimp]

Ah, I see now. I missed that huge vgnd label at the bottom of the triangle. Doh!

Another thing you might want to try - and you maybe can do this without removing the board - is parallel a 10k resistor across each of the 22k resistors in the mixer circuit (2nd opamp stage) this will give you more gain in that stage. Start with the trimpots set for their max resistance, then as you decrease their resistance, the gain of the circuit increases (again, listen for distortion in the reverb return).

If you've got a gain of 25 in the last stage before thr el34, then I am thinking there is no reason this shouldn't work.

Yes, if you have the opamp's inputs reversed, it won't work - but your schemtic is correct: non-inverting input to vgnd, output to inverting input through 150k, and outputs of brick to inverting input through trimpots + 22k resistors.

A standard regulator (7805) requires a minimum of 2V of headroom. So at 7V, you're right on the edge. An LDO regulator only requires about 0.4V of headroom, so it will regulate all the way down to 5.4V at the input.

I don't think the dry signal on the other side of the 220k resistor is getting in the way.

[Theashe]

Could you use your 7.5V tap off of the 15V regulator as an input to your L7805? Might be able to scrap a few components that way. L7815s will regulate up to 34V and can handle 1A, so you can give them a little extra.

I've used 7815/7915 regulators in a 15V split rail supply drawing 300mA from a 500mA transformer and haven't noticed any sag in voltage.