SE EL34 Plexi-verb

[JazzGuitarGimp]

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.

Unfortunately, no. The 7.5V rail is derived from a resistor divider network of two 1K resistors. The Belton reverb brick needs a minimum of 60mA of +5V current. That much load on the 7.5V rail will pull the rail down to some lower voltage.

[Theashe]

Ah, yes, you're right Lou. Let me toss another idea at you.

A quick look at the datasheet says that the L7805 can regulate up to 25V and 1.5A. I'm not sure how high of an input voltage is being regulated to 15V, but if the L7805 input was connected either in parallel with the L7815 input or connected to a broken down zener to create a lower voltage node, it might still do a good job regulating to 5V. Just thinking out loud here, sorry if I'm derailing the original problem.

[JazzGuitarGimp]

Ah, yes, you're right Lou. Let me toss another idea at you.

A quick look at the datasheet says that the L7805 can regulate up to 25V and 1.5A. I'm not sure how high of an input voltage is being regulated to 15V, but if the L7805 input was connected either in parallel with the L7815 input or connected to a broken down zener to create a lower voltage node, it might still do a good job regulating to 5V. Just thinking out loud here, sorry if I'm derailing the original problem.

Hi Theashe,

This poses two other potential problems.

- In this case, it's a moot point, because the current is so low, but let's look at a 5V load that draws 0.250A. If we have 25V on the input of the 5V regulator, then the regulator has to dissipate (25-5) x 0.25 watts as heat. That's 5W of heat. But even at 60mA, that's still 1.2W of heat, which in a guitar amp chassis (already hot), means a heatsink will be required.

- The +21V rail available in this design is made from a 5W filament winding using a voltage tripler. With a voltage multiplier circuit like this, the more current you draw, the bigger the electrolytic capacitors in the multiplier need to be to keep the output voltage from sagging to unusable levels.

Cheers,
Lou

[Theashe]

- In this case, it's a moot point, because the current is so low, but let's look at a 5V load that draws 0.250A. If we have 25V on the input of the 5V regulator, then the regulator has to dissipate (25-5) x 0.25 watts as heat. That's 5W of heat. But even at 60mA, that's still 1.2W of heat, which in a guitar amp chassis (already hot), means a heatsink will be required.

- The +21V rail available in this design is made from a 5W filament winding using a voltage tripler. With a voltage multiplier circuit like this, the more current you draw, the bigger the electrolytic capacitors in the multiplier need to be to keep the output voltage from sagging to unusable levels.

Cheers,
Lou

Hey Lou,

Thanks for indulging me. Two more ideas to solve those problems. For item 1, suppose you placed three 3.3V zener diodes in breakdown and a large resistor to ground at the input of the L7815, and then connect the L7805 input at the junction between the resistor and the third zener. Each zener would dissipate 0.2 watts of heat (0.5 watt zeners are small and easily available). The regulator would then regulate 11 volts to 5 volts, which dissipates 0.36 watts with no chance of drop out.

As for the current draw sagging the capacitors, I think 4700uF should be sufficient, no? If not, 10000uF/35V capacitors can be found pretty easily and cheaply.

Thanks again. Happy Wednesday!

Dave

[JazzGuitarGimp]

Hi Dave,

That sounds doable..... Except if you're laying out a pc board with the intent of selling them to the public. At that point, three or four 10,000uF 35V caps make the board REALLY BIG.

But seriously, if you're willing to use a low drop out regulator, then the 5V supply can be done right off the 5V winding, leaving the multiplier to only worry about supplying enough current for the dual opamp circuitry.

Cheers,
Lou

[Theashe]

I won't lie, I'll be attempting this same project in a few weeks for my next amp, and I don't have a 5V winding on my transformer. I'm using a 24vct for a 15v split rail PS and pondering the best way to get 5v tapped off.

[JazzGuitarGimp]

Dave,

I don't mind continuing this discussion, but I think we have derailed Phil's thread long enough. May I suggest you either send me a PM, or start a new thread for your project?

Thanks,
Lou

[Phil_S]

FWIW, I don't mind the sidebar on power supplies. It's part of the project.

I have an update, but no news. I found pin 5 of the op amp was connected to nothing. I made a jumper to the VGND and verified continuity. It made no difference. I also set the level mixer trim pots to as close to 0 as they will go, just few Ohms. It is too daunting to parallel resistors across the pair of 22K without pulling the board again and I'm not sure there is a practical way to do that.

VGND is 7.3V. I did not have 1% resistors to use in the divider. Although they both metered very much the same, I guess heat is a villain.

I ran a guitar into the amp. Then I pulled the brick. No change, so reverb is zero.

How do I determine if the I've blown the op amp and/or the BDTR-3?

BTW, on the main stage, this turned out to be a very good sounding amp! Rolled back a little, it's very chimey. Turned up, it has plenty of grit. The controls seem to work OK. MV is wired backwards

[Theashe]

Phil,

Just an idea here. The BTDR signal and power ground are connected to 0 volts, while the input buffer is floating on 7.5V, correct?

Without a cap between the buffer output and the input of the BTDR-3, your input signal is still floating on 7.5 volts, higher than the 5V supply voltage for the BTDR. Maybe that's causing the problem? Same goes with the output of the unit, since the DC offset will be different going into the recovery amplifier. Someone else should double check my idea, because the datasheet says that the inputs to the BTDR are capacitor coupled, and I may be wrong because of that.

[JazzGuitarGimp]

Phil, do you have a signal generator and/or oscilloscope?

[JazzGuitarGimp]

Phil,

Just an idea here. The BTDR signal and power ground are connected to 0 volts, while the input buffer is floating on 7.5V, correct?

Without a cap between the buffer output and the input of the BTDR-3, your input signal is still floating on 7.5 volts, higher than the 5V supply voltage for the BTDR. Maybe that's causing the problem? Same goes with the output of the unit, since the DC offset will be different going into the recovery amplifier. Someone else should double check my idea, because the datasheet says that the inputs to the BTDR are capacitor coupled, and I may be wrong because of that.

The BTDR-3H has coupling caps for its input and outputs built-in, according to the datasheet.

[JazzGuitarGimp]

Dave,

Continuing with Phil's blessing...

With a 24VCT transformer, there wont be enough head voltage available to use 7x15 regulators. You'll get (12 x 1.414) - 0.7 = 16.27V but to use 7815 and 7915, you'll need a minimum of 17V to keep the output in regulation. Keep in mind, also, that under low line condidtions, which is usually defined as 120V - 15%, the head voltage could drop as low as 16.27 x 0.85 = 13.83V. You could use LDO regulators (but it may still fall out of regulation under low line conditions), or you could do it without regulators - just use filter caps large enough to keep the ripple low, say under 50mV p-p.

A sure-fire solution would be to use a 32VCT transformer. This would allow you to use 7x15 regulators, and it would stay in regulation under low line conditions.

For the 5V supply, I would use a separate rectifier (two diode, full wave) and its own filter cap > zener string > regulator - or - omit the zeners and use a heat sink on the 5V regulator.

Cheers,
Lou

[Phil_S]

I have a scope. I have only used it once or twice. It's an oldie but goodie, 2 channel 15Mhz B+K 1472C. AFAIK, it works fine, however, may not be calibrated. Calibration probably doesn't matter for our needs. There is a 1KHz test signal I can take from the scope into the input jack of the amp. I made a cable to do this.

I might have a signal generator. I have an old laptop (Win 95, I think) that I loaded a freeware sig-gen program onto. (Quaint...I had to put the sig gen on a CD to move it to this computer.) I will need to build a cable from the headphone out. I have a small pile of old earbuds I've collected for the day when I need to do this. If you think 1K won't do the job, I'll see if I can crank this up.

[Theashe]

Lou - thanks for entertaining my idea. The transformer primary is 115 and line voltages are 125 in my area, so that puts me up around 18.4 volts. I'm going to be smoothing with 10000uF caps. You're right, though, regulation may not be necessary - or LDO regulators are looking a little more attractive because of the higher voltages. I've never used LDO regulators before to experiment with. What are your preferred ones for +/-15V? I use L7815/L7915 because I can buy 20 of them for 2 bucks lol.

Phil - You can use your cell phone as a portable signal generator - just download a tone generator app and splice together some gator clips on a 3.5mm cable. Or, your scope may have a 5Vpp square wave generator on it, which may be good enough to trace where the signal disappears to.

[JazzGuitarGimp]

Phil, the 1KHz, 1Vp-p square wave calibration output on the scope should do nicely.

Vgnd of 7.3V is fine.

The brick is in a socket, and can be removed, yes? If so, take it out. Disconnect the input wire to the reverb circuit at the 220k / 470k junction, and input the 1KHz signal into the buffer opamp. Turn the 1M input trimmer all the way up. Set your scope for 200mV per division, with the input mode switch on AC Coupled. Set the timebase to 200nSec per division. Start by confirming the square wave at the input of the board is actually 1Vp-p - the square wave should stand 5 divisions tall. Don't worry if this is off by a bit. If it looks like only half a division tall, then you are using a x10 probe. See if the probe has a X10 / x1 switch, and set it to x1 if it does. If it doesn't, the set input to 20mV per division. Also, make sure the variable input knob is set to the CAL position (fully CW).

Assuming that looks good, move the scope probe to the output of the buffer (pin 1 of the opamp). You should see the same signal there, both in shape and amplitude.

If that goes as planned, input the 1KHz square wave into one of the two inputs of the revovery opamp. Do this at the brick socket, so that you have the trimp pot and 22K resistor in the signal path. Set the trimmer for maximum resistance, so that you've got aprox 42K ohms between the generator and pin 6 of the opamp. Set the scope for 500mV per division (or 50mV if using a x10 probe) and probe the ouput of the recovery opamp. The square wave should stand just over 7 divisions tall.

Report back with your findings,

Cheers,
Lou