Neovibe Debug Help

[R.G.]

As always - try with what you have. It may work until you find the perfect thing.

And there is some possibility it may work *better*. That's how a lot of effects got "discovered".

[Jaicen_solo]

Glad this thread has been resurrected again.

I was about to post but R.G. beat me to it. I would have gone for the lowest mA bulb you can, just get it working and go from there.
I'm not sure where you're located, but Steve at Smallbear can provide you with a bulb and LDR that works perfectly for the Neovibe. I built mine using locally sourced parts, and it was never right till I gave in and ordered from Steve. Now it's perfect, absolutely perfect. I added a symmetry control to the bulb so I can have the choppy Machine Gun sound from Band of Gypsies, or the more subtle early leslie type sound.

I said I was happy about the resurrection, as I was going to ask R.G about the power transistors. Which one do you recommend for mounting on an external heatsink? I'm still not happy with the temperature of the encapsulated transistors i'm currently using.

[R.G.]

Actually, anything with an exposed metal tab is dramatically better than a TO-92 for getting heat out. I really like the 3902 and the lower voltage version. The MJE270 is great. I guess I'd say that the TIP110 is my least favorite, but in the bigger package, it will dump heat better with no heat sink. A small square of aluminum clamped to the tab of the transistor just about sews things up. I suspect that the heat sink square is not even necessary.

The driver temp thing is a weak point of the original design. I repaired a number of original Univibes whose only flaw was a burned out driver transistor. Even then, the majority of them worked fine semi-forever with the original device. It might be that epoxying a metal square to the flat side of a TO-92 would be enough.

A TO-92 typically has a thermal resistance to ambient of 125 C/W - meaning that if you dissipate 1W in the transistor and it just sticks up into earth-normal air, the chip rises 125C over the outside air temperature. Most TO-92 devices set a maximum of 0.625W to keep the chips below destruction temperature.The TO-126/TO-225 case has a thermal resistance of 83C/W with no heat sink, and a thermal rating of 1.5W in free air. That step is enough to get you over to the "probably never fail" side of things. The TO-220 package usually has a thermal resistance of 62C/W, and a dissipation of 2W in free air. Both the 126 and 220 have thermal resistances to a heat sink of only a few C/W, and so their power capability jumps dramatically with any heat sink on the metal tab at all.

The trick in the neovibe lamp driver is getting a device that has enough current gain, and can dissipate a watt or so reliably.

Well, there are a couple of other tricks. The power in the driver is equal to the voltage across it times the current through it. The current is the lamp current, of course; the voltage is whatever voltage is left over after the lamp does its thing. So if you have a 50ma *6V* lamp, the lamp lights up most with only 6V across it and 50ma. The transistor is eating 9V*0.05A = 0.45W on peaks. Still do-able, but a stretch for a TO-92. Much better to have a higher voltage or lower current, or both in the bulb. A 12V rated bulb reaches rated current at 12V, leaving only 3V across the transistor at max current. This cuts the power in the transistor dramatically, and moves the max-power dissipation point from the peak brightness on the bulb down to mid-brightness, where the bulb and transistor have more nearly equal power dissipated. If you have a low-voltage bulb, putting in something to soak up some of the excess voltage the bulb doesn't need helps a lot. This is where inserting an LED for indication helps the driver. A zener or resistor does the same thing for the driver. These only work if you have the voltage to spare, of course.

[Jaicen_solo]

Thanks R.G.

[darlon]

Hi all

so after all I finished my neovibe (almost)

I used:
BC550 as a lamp driver, but removed R50 as I first thought that lamp is not shining very bright. used 22 mm brass pipe cap as a shield - first tried blank and next with black isolation tape inside. IMHO no big difference but I think that tape is gonna stay there after all.

used 12V 40 mA bulb - not bad, actually quite a big surprise that it was very bright at the peak. I'm not quite sure if that will be better if lap is brighter or darker in peak but it is sounding ok with bias pot maxed towards GND. at maximum depth it gets a little wobbly, but turning the pot a little down does the job fine - it is always better to have some margin.

LDRs facing the bulb - i decided to set them up almost touching the bulb, don't know if it's good but not sounding very bad.

I was very surpiresd also that lamp driver is not getting any heat at all. left the board flashing for one hour and BC550 is not warm, maybe has like 5 Celsius degree more than other 2N5088's!

Comparing to factory made uni-vibe made by dunlop its way more cool I think.

It also seems to me like the effect is a little bit dark, lacking a little high freq. when adding more depyh its getting better - any idea why? maybe the bulb is too dark?