It didn't catch fire . . . now with questions !

[paul.creedy]

. . . but I did blow a hole in an LM386

Working my way through a few simple circuits before attempting anything major, I built a Tone God Finish Line and finished it last night.

It works, but doesn't do much other than fuzz so I'm not getting all the exciting things promised

So, having swapped both ICs and transistors successfully and with all three in sockets, I changed the first LM386 and swapped a pair of 2N5089s for 2N3904s (hFE of 212 instead of 490).

Those who know much, much more than me might well consider that idiotic, but I'm still closer to "lets see what happens, that kind of worked with another circuit, maybe something interesting will happen" than "I know enough to know that doing that might not be a good idea".

When I first turned the pedal on it worked, as I said, but was very quiet, even with the Drive and Vol close to maximum. This time there was no sound at all once I turned it up (I always start with the volume right down in case of horrific noises) and when I turned it off there was more silence for a few seconds and then a pop.

So while I'm sorry to tell you that nothing burst into flames, I did break my first component

I'll do a proper debug next week when I have a moment and report my findings if it's still refusing to play nice.

[ElectricDruid]

Good work. Welcome to the club

Carry on, sir.

Tom

[paul.creedy]

Everything for a reason

As a consequence I've learned that there are three types of LM386 (N-1, N-3 and N-4)*, which I didn't know earlier.

I've got 10 that are marked up as LM386 M-82 (which may or may not be LM386N-4) but for experimentation purposes I shall order some "proper" ones from Bitbox (all the others I have came from eBay in a pretty random fashion).



*I include these facts for those like me who know almost nothing.

[willienillie]

M-82

That would be the Barrett M-82.  Quite a pop indeed.

[PRR]

> there are three types of LM386

I was thinking: two ways to blow a '386. Power backwards (even that may not be fatal), or excess voltage.

There's very little reason to feed over 9V except maybe for a headphone amp.

[ElectricDruid]

there are three types of LM386

Warm, hot, and smoking?

T.

[pinkjimiphoton]

welcome to the magick smoke club!

we're all charter members

[paul.creedy]

As may be clear from my posts so far, I'm not starting my pedal building explorations with a sound knowledge of electronic circuits and the mathematics behind them - I did consider this (undoubtedly correct) approach but despite owning books which would teach me the basics, I have found the approach of starting with simple circuits and learning as I go more engaging, even though I'm sure it's far from ideal.

With this circuit I have reached a point where doing a little research/learning before continuing appears sensible, so I'm going to investigate how the LM386 works within the circuit and also how to bias a transistor (and why), rather than continue with my "swap transistors with wildly different values to see what happens" approach.

Here's a link to the vero layout I've used - https://tagboardeffects.blogspot.com/2014/01/the-tone-god-finish-line.html

1. I discovered that the LM386 gives a gain of 26dB, but if pins 1&8 are connected with a 10uF cap it will increase that to 46dB. Currently (not debugged) the circuit is very quiet, if I wanted to add more oomph and connected 1&8 would that work or would it throw the whole circuit into chaos ?

2. The layout indicates wiring the input to tag 3 on the "drive" pot rather than 2 which seems the usual method, however I'm finding that when the gain is turned up full the signal drops out completely - could the two be connected ?

3. At the time I finished the board I didn't have any 1N5817 diodes, but as it says it's there for polarity protection I added a link instead, was I wrong to do so ?

4. I've got a simple component tester and a digital multimeter, will that do me for now or should I consider investing in an oscillator sometime soon?

[GibsonGM]

1) If the design didn't call for the additional gain, it's probably not going to sound good, but you can try by tacking wires at each pin and wiring in the cap 'offboard' temporarily to see what it sounds like.   I predict feedback and noise.

2) might be fixed by adding a very small resistance to the end of the pot, maybe 100 ohms, to prevent using that 'last gritty bit' of the pot when you connect wiper to outer lug.   

3) The wire is necessary for this to work, we can see that...but you have no protection.    Better than a series diode like that (which steals .5 volt or so from the circuit) is one that is reverse-biased between power and ground, and easy to add after.   Virtually any diode will do.  1N914, or emitter/base junction of a BJT....but ideally you should have some 1N4xx diodes on hand for this and power supply work. 

4) a resonably well-made DMM is fine for quite a while in the work we do. No need to rush out to buy and O-scope!  Better is to save up and get a NICE DMM, something very reliable in the $30 to $50 range, perhaps, if you continue with the hobby.       It will help you do better work.

[paul.creedy]

Thanks for that

I shall go and do some debugging, voltage checking and tinkering in a while, and return with conclusions later

[bluebunny]

. . .  if you continue with the hobby . . .

Ha ha ha ha ha ha ha ha ha ha ha!  No, stop, Mike!  Please, I wet myself a little...         

[paul.creedy]

Right, I have checked everything is in the right place and have the voltages. I started with it as it was (I'd put BC108s in with hFEs of 426) and later swapped the 108s for 2n5089s with hFEs of 496. I was using an EHX power supply which was sending 9.82v to the board and other than that nothing was plugged in.

The 8 points of the LM386 read:
1. 1.30v
2. 0.0
3. 0.0
4. 0.0
5. 4.77v
6. 9.81v
7. 4.99v
8. 1.30

Q1 reads:
C. 4.70v (4.49 with the 2N5089)
B. 4.78v (same)
E. 4.78v (4.43 w/5089)

Q2 reads:
C. 4.70v (4.49)
B. 4.78v (4.78)
E. 7.33v (4.43)

I've got other voltages if needed.

[Kevin Mitchell]

welcome to the magick smoke club!

we're all charter members

There should be a thread for this club 


Sorry... Carry on.

[GibsonGM]

. . .  if you continue with the hobby . . .

Ha ha ha ha ha ha ha ha ha ha ha!  No, stop, Mike!  Please, I wet myself a little...         

I mean.....SOME must escape addiction, no?  Or...is it HOPELESS?   

[paul.creedy]

here's a picture of mine

[duck_arse]

4 types of LM386; the three you listed, plus the ventilated type.

I can't work out what the collectors on that vero are supposed to be doing - any chance you can post the schematic for this?

[paul.creedy]

4 types of LM386; the three you listed, plus the ventilated type.

I can't work out what the collectors on that vero are supposed to be doing - any chance you can post the schematic for this?

aha!

With that question (and my finding and looking at the schematic) I do believe you have solved the problem

My brain, such as it is, looked at the requirement for one 2N5088 (which I have) and 2N5087 (which I don't) and thought "maybe a 2N5089 will do", not realising that the 2N5087 is PNP, not NPN.

I'm not saying the transistor numbering system is mad, but it's clearly confusing enough for a newbie to make a daft decision. Never mind, I'll continue to live and learn !

I shall now order some 2N5087s (or equivalent) and once the circuit is correct, see if it works

[paul.creedy]

. . . not realising that the 2N5087 is PNP, not NPN.

Those more knowledgeable might have replied at that moment "the 2N5086 is PNP too . . . " to which I would say, after checking my PNP box "ooh, I've got some of those".

With one fitted, along with an NPN suitably similar in hFE, the pedal is much closer to what I was expecting

As there are more LM386s on the way, along with some 2N5087s I've just ordered, it'll be a while before I consider it completed, but I'm a lot closer to that than I was this morning.

Thank you all once again for your help  

[duck_arse]

that's all well and good, but still no sight of the circuit diagram.

ohh, and BC557, 558 and 559 are all PNP. and BC177, 178, 179. the general rule is - whatever you have at hand, put it in and see. worry about correcter parts later.

[paul.creedy]

that's all well and good, but still no sight of the circuit diagram.

there you go - isn't google a wonderful thing ?