Restore/Mod an old BBD Echo

[Rutger]

Hi,

I just bought a cheap old Phonic Echorder PH8020. It's a nice echo, but there are 2 things I would like to fix:

- There appears a soft, high beep when I turn the 'delay' knob above 10 o clock. I remember this beep from a new BBD echo which I had a couple of years ago, and remember that I fixed it with the help of a tech. It had something to do with calibrating trimpots.
- The delay time is too sort for some styles. Slap back is great but for U2 kind of playing it's just a little too sort. I wonder if I can alter the delaytime somehow.

I would like to do it myself, but how? I can't find a schematich for it. From what I can find, this unit is brought on the market by several unknown brands (Phonic, Better, Gemini, Maplin, etc).
I can emagine this could be a copy of some well known echo, so probably I can use another schematic and find my way through it.

So, can you help me out somehow?

It's identical to this one, except for the colour and knobs:

[Mark Hammer]

Check the board and see if it says "Melos" I have acouple of older MN3005-based delays that have the exact same control/switch/input jack complement.

I'm guessing this one also uses a single MN3005, and does not have companding (input attenuation and a clipping indicator LED is usually a sign).

Virtually all BBDs employ two parallel sample-shifting pathways, internally. The chip has two actions to perform: take a sample, and move it along. A complementary clock signal (that I will call "tick" and "tock" here) feeds two pins on the BBD. While tick is getting one side to take a sample, it is making the other side pass its sample along. When tock occurs, that second side takes a sample, and the first side is now moving its sample along. The two parallel signal pathways are eventually combined at the output, and their respective samples are stitched together to make one continuous sample of the input signal.

What one will sometimes see at the two outputs is a balance trimpot (not the same as the "Balance" control on the front panel). Because tick is high when tock is low, and vice versa, if the two BBD outputs are "perfectly" balanced, then the bits of tick and tock in the audio path cancel each other out, and the whining one hears is reduced. If the circuit includes companding, lots of steep lowpass filtering, or both, that whining is made even less audible.

But all that extraneous-noise reduction begins with adjusting the balance trimpot until the audible clock whining is minimized.

There MAY be a trimpot on the board that fine tunes the min/max delay time. Keep in mind that the lowpass filtering in the circuit aims for the widest possible bandwidth with the least noise. If the filtering is set for a 10khz corner frequency, but the chip is being told to clock slowly enough to produce 500msec delay, the clock signal will actually be below the filter's cutoff frequency. Some analog delays, like the old MXR and Ross units, had tracking filters that would adjust the cutoff frequency as the delay time was changed, but not the unit you show. So there can be a discrepancy between the clock noise that the filtering and overall design anticipates cleaning up, and what the unit actually produces.

This is the long way of saying that if there is a trimpot for adjusting overall delay/clock range, sacrificing a bit of delay time will also help cure the whining.

[Mark Hammer]

Also, the more delay stages an audio signal passes through, the more degraded it gets; especially if the delay time is longer and the clocking slower. Each repeat that is produced through the feedback/repeat loop ends up passing through 4096 more delay stages. If you set the unit for multiple repeats, they will exaggerate the noise and be unpleasant.

What I do is attached a small-value cap from the output of the repeat pot to ground. This has a very shallow filtering action, and will remove a little bit more treble and upper mids from the delay signal with each repeat, so that successive repeats become more and more dull-sounding. This will NOT affect the first repeat, and is separate from the fixed filtering imposed on the overall delay path; it is exclusively targetting the feedback loop.

I find this adjustment not only helps to clean up noise, but makes the additional repeats sound more natural. Remember that, in the real world, reverberant surfaces are imperfect, such that late reflections have much less high-frequency content than early reflections. This mod mimics that effect to some extent.

What is a "small" value? Depends on the pot value, I suppose, and the degree of noise. I would start with something in the .01-.047uf range, and see how that works for you.

[Rutger]

Thanks Mark for the clear explanation! Yesterday I took a quick look at the inside. Didn't have the time to pull out the board (it's placed upside down) so I couldn't check for the brand, but I found the balance trimpot. I turned it both ways but I guess it was already set at the best spot. But I think it's some kind of grounding problem, for the noise reduced and the beep nearly vanished when I touched the chassis/trimpot. Does this mean your 'small value cap' will be the solution?

I hope I've time this evening to check the board for another trimpot for the delaytime. But if there isn't any: how is the delaytime set, and can I change it somehow (e.g. by replacing a few resistors)?

[Mark Hammer]

There will probably be an MN3101 clock chip in there, with a small-value capacitor (usually less than 1000pf) close to it. Making that cap value smaller will increase the clock range and reduce the delay time, and making the cap larger/higher in value does the opposite.

That's one way to alter max delay time, but probably too general, rather than finely tuned. The better way would be to find the trimpot for adjusting clock range...assuming there is one. If there IS one, then set the delay time on the front panel to longest delay (hence, most "whine") and adjust the trimpot until the whining goes away.

The smartest move, however, is to start with the balance trimpot I mentioned earlier (and as you have already done), and cancel as much clock noise as you can first, and THEN use the more heavy-handed technique of altering the max delay time.

The suggested additional cap-to-ground is for the feedback path only, and will only affect any additional noise produced by recycling the delay signal back through the BBD again and again. It is still possible for there to be audible whining with NO additional repeats if the clock range is set too low for the lowpass filters to filter out. In that case, find the cap connected to the clock, and make it a bit lower/smaller.

An easy experiment to do would be to locate the pads on the copper side of the board that the existing cap is connected to and temporarily add more capcitance to ]lower the clock range (and increase the delay time). If this makes the whine even more annoying, then clearly the problem is as I described: mismatch between the clock range and what the filters were designed to block out.

[Rutger]

Thanks again Mark! Yesterday I experimented with the balance trimpot again and it is set at the best spot now. Still some noise and beeping, but found out that there was a groundproblem: the frontpanel needs to be grounded, it makes a great difference in noise and the 'beep' is now a lot less noticable. Now I can turn the delay-knob to 3 o clock without this noise.

Next step is altering the delaytime. I haven't found another trimpot yet, so there probably isn't any. Unluckilly the whole unit needs to be taken apart to get the board out and take a good look at it. So I will come to that when I have some more time left, and experiment with the different capacitances as well.

But the unit is already a lot more usable now!

[Mark Hammer]

We had to put up with a LOT in the early days of analog delay, but we were just so happy to actually HAVE delay available in portable form at reasonable cost, that we overlooked all the weaknesses.

Now that the available technology has improved so much, the standard is set much higher, and all the things we used to accept willingly stick out as major flaws or "problems".

[Rutger]

One more question about this echounit. It can be controlled by a footswitch. Do you know if I need a momentary or a latching switch?

[Mark Hammer]

The switching in these is nearly always electronic, using a FET to lift the delay signal (the dry audio path is always in circuit). If it used a momentary switch, there would need to be some sort of flip-flop to change states, and stay latched in a state. That would either require a 4013 or 4027 or 4007 CMOS chip, or a discrete circuit like Boss uses. Much easier and cheaper to use a latching footswitch that directly turns the FET on or off.

So I'm going to go with latching switch.

[Rutger]

Thanks!