A little help with a Neovibe?

[skrunk]

Almost finished populating a neovibe pcb (ver. J 10/07).
I was going to put in all 2N5088 for the trannys then read something about using MPSA13s for all but the first 4 trans. and the lamp trans (q13, which I might try a TO-220 type).

Is there much to be gained by using the darlingtons?
I don't even have MPSA13s, but they are MPSA13Gs.
Should I go with those, try find some MPSA13s, or just go ahead with the 2N5088?

Bread-boarding this thing is beyond my skills and I'd like to avoid using half a dozen sockets.

I also did the unity gain mod by changing R8 and R9 to 894R (closest I had) and 3.6k respectively and noticed RGs March 2010 update about the input mixer.
Would putting a 2.2M in for R3 along with the above changes compensate for any level drop?

Any advice would be gratefully received.

[PRR]

> read something about using MPSA13s

Be careful what you read. Whatever it was.

Look at how, for instance, Q4 Q5 are connected together. This can be _two_ 2N5088 *OR* _one_ Darlington such as MPSA13. Same for the Q6+Q7 and Q8+Q9 pairs. Also the Q11+Q12 pair. It comes out EXACTly the same (for our purposes).

If you thought you read to put _TWO_ MPSA13 in each stage (one per 2N5088 hole).... that does not seem right.

In mass production, and some super-teeny circuits, the single pre-made Darlington is a penny cheaper and a tenth-inch smaller. Sometimes that's enough advantage to use the MPSA13.

For DIY on a pre-made PCB with 5088s in hand, just do it like it shows, two 5088 at each stage.

[skrunk]

thanks PRR, much obliged!
yeah I was a bit confused about that, my limited understanding of a 'darlington pair' was 2 trannys connected from say B to E.
talk of mpsa13s kind of threw me a bit! I think it referred to a different layout.
I'll stick with the 2N5088s then.

has anyone tried RG's suggestion of using a <2.2M at the input mixer with or without the R8/R9 gain mod?

[skrunk]

a few quick questions if I may.
I'm about to try mounting the film cannister and bulb/optos.

I have a black film cannister with a black lid.
I noticed the instructions call for a grey lid.

Will a black lid do?

Should I glue the cannister to the pcb or will the soldered parts hold it in place?

I got 3 different bulbs from smallbear.
The big one with the screw in socket (1.5v/25ma). (think I got the wrong socket, mounting holes don't match pcb).
and the small one (1.5v/25ma) with the red and white leads coming from it (don't know how I'd mount that)
seems like these 2 are too small a voltage?
The 12v/40ma one with the white bulb holder (lead spacing is a hair too wide (0.5mm) but should fit ok if not too secure, looks promising).

Any thoughts on which one is best to use?

also there are 2 sets of holes for each opto, should I use the ones nearest the bulb or the other ones?

any general tips for this procedure or photos of completed builds would be greatly appreciated!

[jasperoosthoek]

Watch out for using too high gain transistors in the input buffer (first three at the input). Mine oscillated when I used transistors with a gain of about Hfe=280. I fixed it by using 2n3904s with a gain of 190.

[RedHouse]

> read something about using MPSA13s

Be careful what you read. Whatever it was.

Look at how, for instance, Q4 Q5 are connected together. This can be _two_ 2N5088 *OR* _one_ Darlington such as MPSA13. Same for the Q6+Q7 and Q8+Q9 pairs. Also the Q11+Q12 pair. It comes out EXACTly the same (for our purposes).

If you thought you read to put _TWO_ MPSA13 in each stage (one per 2N5088 hole).... that does not seem right.

In mass production, and some super-teeny circuits, the single pre-made Darlington is a penny cheaper and a tenth-inch smaller. Sometimes that's enough advantage to use the MPSA13.

For DIY on a pre-made PCB with 5088s in hand, just do it like it shows, two 5088 at each stage.

He's probably referring to the Q10 and LFO darlington mods from the Forum-Vibe.
(see:  http://classicamplification.net/forumvibe/ForumVibe.pdf)

[R.G.]

I have a black film cannister with a black lid. I noticed the instructions call for a grey lid. Will a black lid do?

Black will be fine. All the film canisters I had seen for a long time had gray lids, but black is fine. Just so it's opaque and does not let in light.

Should I glue the cannister to the pcb or will the soldered parts hold it in place?

I need to go look at the instructions. As I remember it, after modifying the canister part to be shorter and still clip into the lid, I put the lid on the PCB, top/flat side down, so I could press the shortened canister onto the lid.  I believe I stuck the lid down with double sided tape so that the canister would not stress the solder joints, but there's probably not a huge difference. Probably 90% of the lifetime stress is during the debug/tinkering phase.

I actually prefer the metal "house" approach personally. But both work.

I got 3 different bulbs from smallbear. The big one with the screw in socket (1.5v/25ma). (think I got the wrong socket, mounting holes don't match pcb).
and the small one (1.5v/25ma) with the red and white leads coming from it (don't know how I'd mount that)
seems like these 2 are too small a voltage?  The 12v/40ma one with the white bulb holder (lead spacing is a hair too wide (0.5mm) but should fit ok if not too secure, looks promising). Any thoughts on which one is best to use?

Probably all would work, but I'd start with the 12V/40ma one.

also there are 2 sets of holes for each opto, should I use the ones nearest the bulb or the other ones?

I put multiple sets of holes to fit different sized optos. Just use the ones which fit your optos most easily and solidly. Use the holes that put all the optos the same distance from the bulb.

any general tips for this procedure or photos of completed builds would be greatly appreciated!

There are several people's pictures of their Neovibe boards here in the forum.

[R.G.]

has anyone tried RG's suggestion of using a <2.2M at the input mixer with or without the R8/R9 gain mod?

I breadboarded the input preamp and tried it, but I don't have a working Neovibe at the moment to try the effect end to end.

Both in simulation and in the breadboard, the output was as expected for the stock and modified input. It's not too tough to put the higher resistors at the input and then clip them out to replace with the original values if the sound doesn't please you. Some people would prefer the "browner" or "duller" (depending on how you look at it) sound of the extra loading. Others will like more sparkle.

[jasperoosthoek]

@R.G.

Did you ever get any oscillation of the input stage? It was a square wave at about 300kHz, peak to peak amplitude of maybe 7 volts (don't exactly remember how much but it was big. I got it on the PCB ordered from from GGG. Lower gain transistors solved the problem for me.

[skrunk]

thanks for the responses 
never thought of putting the lid at the bottom,
it would surely be easier so I'll try that.
the metal box option is just beyond my means.

Watch out for using too high gain transistors in the input buffer (first three at the input). Mine oscillated when I used transistors with a gain of about Hfe=280. I fixed it by using 2n3904s with a gain of 190.

didn't want to hear that, my 2n5088s are all about 480 hfe.
I got some low gain 2N3904s here though, maybe I'll try those...

has anyone tried RG's suggestion of using a <2.2M at the input mixer with or without the R8/R9 gain mod?

I breadboarded the input preamp and tried it, but I don't have a working Neovibe at the moment to try the effect end to end.

Both in simulation and in the breadboard, the output was as expected for the stock and modified input. It's not too tough to put the higher resistors at the input and then clip them out to replace with the original values if the sound doesn't please you. Some people would prefer the "browner" or "duller" (depending on how you look at it) sound of the extra loading. Others will like more sparkle.

thanks RG, I'll try it out.

[skrunk]

ack!! som bitch 1000uf electros are too tall for my 1790NS by about 2mm 

[Jay]

the metal box option is just beyond my means.

It's easy.

I''ve built a couple using metal foil containers, such as you might find an oven ready chicken in at a supermarket.  They have a flat area (did in my case) and without ribs big enough that you can cut out (strong scissors will do) and fold into an open box shape.  You can then glue, tie, tape or whatever to the board.

Practice on a sheet of paper first so you know how big it has to be and how to fold it.  I think I cut a large square, then cut a little snip in each side near the corner to make tabs that folded inside.  You can do all this whilst the chicken is cooking.

I've had to repair a couple of original Univibes, they both had the ldr's soldered flat to the board and the shiny metal box acts as a light scattering mirror so they are equally lit by the bulb.  No need to play around with angling the ldr's toward the bulb.


Jay

[RedHouse]

the shiny metal box acts as a light scattering mirror so they are equally lit by the bulb.  No need to play around with angling the ldr's toward the bulb.

That really depends on your photocells and their sensitivity whether you angle or not.
(in my experience)

Another place to get shiney metal is from old Tuners and Video equipment, they often have a nice little tinned metal box on their RF from ends, you can usually snag something and make use of it.
(and it's solderable)

[Jay]

I just used the epoxy cased cells from Steve at Small Bear.  The original 'vibes I worked on (two slightly different board versions, one with carbon comps the other carbon films) had larger metal cased cells but in both of them they were soldered tight and flat to the board.  No idea as to the characteristics of the cells though compared to Steve's.

Both these old Univibes sounded good, but one was better than the other.  With the lamp driver trimmed and RG's wet/dry mix tweak done mine is identical sounding to the better of the two.  Very satisfying!

[skrunk]

got a quick wiring question,
where do the K and J pads on the pcb (ver J 10/07) go?
I'm looking at the pdf but not sure what to do.

I see they're connected to the bridge rectifier + and - but not sure how they connect to the DC jack.
I'll be using an 18v DC power supply into a Boss style insulated DC jack.

thanks!

[R.G.]

got a quick wiring question,
where do the K and J pads on the pcb (ver J 10/07) go?
I see they're connected to the bridge rectifier + and - but not sure how they connect to the DC jack.
I'll be using an 18v DC power supply into a Boss style insulated DC jack.

It depends on what else the 18V dc power supply feeds. If you don't connect it to **anything** else, you can simply connect it to pins K and J any order.

If you connect the 18V supply to ground on something outside the neovibe, it won't work.That's because the signal ground on the other stuff you connect it to shorts the 18V supply wire to the diode bridge, which is then trying to be "ground" on both sides.

K and J are intended to be where AC comes in. Putting in DC works, because the bridge rectifier makes it be the correct way round for DC, but then neither of the wires connected to K and J can be connected to signal ground, even outside the neovibe.

[skrunk]

thanks RG,
I won't be using the power supply for anything else.
so I'll be connecting the K and J pads the to + on the DC jack?
do I then just connect the - lug of the DC jack to some earth point?
been looking at a few pictures but just can't quite make out how people have been doing it.

[R.G.]

That works fine, then. Just connect K and J to the + and - terminals on the DC jack.
It does not matter which is which, because the diode bridge swaps them around the correct way.

You will lose 1.4V in the diodes in the bridge, so if the DC supply is 18.00V, you will only get 16.6V as input to the 15V regulator. That is very marginal as to whether it will regulate or not. Generally these things need a minimum of 1V across them. Maybe work OK, maybe not. if your nominally 18V power supply is actually 19 or 20, all is well.
You could wire this way up, then measure the voltage before and after the 15V regulator. If you get 18.5V in (for example) and 15V out in actual operation, you're golden.

What you **could** do is to ignore points J and K and wire directly to the hole on the PCB that the diode bridge uses for its + output and - output. That will give you 18V and good regulation. But it will make it critical to get the + lead wired to the hole the + lead of the diode bridge went into and the - lead where the - diode bridge lead went. A good way to do this is to wire up the jack, but leave the wires hanging. Stick the 18V power supply into it, and then measure the voltage on the ends of the wires with your meter before you solder them in. Find out with your meter which is + and which is - and mark them. Then solder the - one into the - hole, the + one in the + hole. It eliminates mistakes if you do it carefully. Don't get this one backwards?

[skrunk]

thanks for your time RG, much appreciated
I plugged the power supply into the dc jack to measure the voltage.
It's a positive tip power supply (electro harm 18v dc/500ma) which if I understand you correctly is ok, as the diode bridge sorts that.
I'm measuring 26.8v at the jack with my DMM.
I wasn't expecting that, sounds like a lot, but I guess it should be ok then.

[R.G.]

thanks for your time RG, much appreciated
I plugged the power supply into the dc jack to measure the voltage.
It's a positive tip power supply (electro harm 18v dc/500ma) which if I understand you correctly is ok, as the diode bridge sorts that.
I'm measuring 26.8v at the jack with my DMM.
I wasn't expecting that, sounds like a lot, but I guess it should be ok then.

What that means is that it sags to 18V when you load it with half an amp. It will sag down toward 18V, but you should still have more than enough to get through two diodes drops lost in the bridge and still have enough to regulate.

That's good. You have enough voltage to just connect it up either way round to pads J and K. The diode bridge will make that come out the right way at the filter caps.