[dgardner]

Glad to see your getting closer..

R19 and R49 need to be 200 ohms. They do not set the gain. Don't change the value. If you used 100 ohms in these locations, the output is not running properly. Slow movement just above and below 0v during operation is normal.

For a gain of 11, set (R16,R46) to 10K, and (R7,R37) to 1K. Make sure (C27, C28) has a 10pF cap in it.

The servo behavior sounds normal, but if your resistors in R19 and R49 are not 200 ohms, I can't predict the results easily. I typically see +/- 2 to 3 mV worst case.

As for the PSU, all parts on the heatsink must have insulators. If you don't insulate, the parts WILL smoke.

 

[CB600F]

Hi there,

<< Finally I started an electronics project again. It has been a while. >>

Has anyone been fiddling around with the dynalo and the SWCAD III Spice simulator? (Or any other spice simulator for that matter)
I have just started with my first simulations and it seems to simulate the amp pretty well. I tried some other transistors like the bc547b; no shocking differences there. I thought the opamp would make a difference, but after adjusting simulation resolution there was no noticeable difference.

Regards,

David

 

[sbelyo]

Does anybody have digikey or mouser part number for:

C6, C7, C19, C20

C1, C2, C3, C4, C12, C13, C14, C15, C16, C17

C10, C11, C21, C22

I just want to make sure they fit and are the right type. I'm starting to gather parts this weekend.

 

[grasshpr]

These are mouser part numbers:

C6, C7, C19, C20 -> 581-SR205C104KAATR
C1-4, C12-17 -> MKT1817410064
C10, C11, C21, C22 -> 5989-100v4.7

 

[bg4533]

Quote:

Originally Posted by grasshpr C6, C7, C19, C20 -> 581-SR205C104KAATR

Minimum order of 3000 on that part

 

[grasshpr]

OOps. Must of changed recently. I had the same part number when I bought a few when I made my dynalo. Sorry, at least you have something to compare with

Thanks for pointing that out bg4533, I surely won't purchase that item again.

Quote:

Originally Posted by bg4533 Minimum order of 3000 on that part

 

[grasshpr]

Here's a similar part from digikey which doesn't have a minimum purchase of 3000!

399-2081-ND

Sorry for the confusion.

 

[dip16amp]

Mouser also has this: 581-SR205C104K and 581-SR205C104KAT

 

[sbelyo]

Thanks guys...

I'm planning on matching the transistors with my multimeter's hfe setting. Beyond that I'm hoping using the 1K pots in place of the 2 499 ohm resistors and switching LED's I should be in good shape.

 

[bg4533]

Quote:

Originally Posted by sbelyo Thanks guys... I'm planning on matching the transistors with my multimeter's hfe setting. Beyond that I'm hoping using the 1K pots in place of the 2 499 ohm resistors and switching LED's I should be in good shape.

I read somewhere here that matching these transistors with your multimeters hfe setting might be worthless. Multimeters generally use less than 1ma to test hfe while the transistors will see between 2ma and 16ma in normal use.

Transistor matching is ideal, but not necessary if you use the trim pots. Just make sure you have transistors from the same class

 

[Pars]

Quote:

Originally Posted by bg4533 I read somewhere here that matching these transistors with your multimeters hfe setting might be worthless. Multimeters generally use less than 1ma to test hfe while the transistors will see between 2ma and 16ma in normal use. Transistor matching is ideal, but not necessary if you use the trim pots. Just make sure you have transistors from the same class

Fortunately these transistors are rather linear across the required Ic range. I have done extensive matching via a test fixture only to have to tweak the resistances/LEDs to get acceptable offset, so I'm certainly not going to say that matching is imperative to get an operable amp. However, I do believe that the better you match stuff, the less you have to deal with down the road and I personally don't trust trim pots... worked too long as a tech in the semiconductor industry to trust them not to drift

.

I think the most rational approach to building a Gilmore-type architecture was that of Stackofhay's (geoff... where is he these days anyhow?). He used a pot initially, but only to get a general resistance value. Then I believe he used a decade box with the pot out to dial in the values required, then put in fixed resistors as close as he could get to those values. Of course everyone does not have a decade box nor tons of fixed precision resistors laying around to make up the right values (or paralleled values). I still think its the right approach however. But then again, I am an idealist...

just my devalued $0.02

Chris

 

[dgardner]

Another thread supporting construction questions on the Dynalo Rev C project is here.

 

[sbelyo]

I thought I might be able to do domething similar by socketing the two pots (or positions) at the top. Then once the offset is adjusted I would remove the pot and take a reading to see what it's value is. Then replace it as close as I can with a rn60d. I only have to order 9 values to cover what's in between 499 Ohm and 1.0K I also just went through my resistor box and pulled out 11 rn60d's that are already needed. So at least my part cost wont go up that much.

 

[sbelyo]

I'm going to build this on some proto-board tonight and then match the transistors that I just matched with my multi meter using the hfe setting.

It's just a schematic without the transistor symbol. I drew it with E for emitter B for base and C for collector. The value I took from gilmore's original project in headwize. I hope it's right

 

[sbelyo]

I built the circuit and I suppose it works. All of the pairs were within 0.01 Volts of each other ( most were a dead match). It gave me the same results as the hfe setting on my meter.

I also went back to the old ones that I couldn't match up and all their numbers were at least 0.10 apart.

Also a side question... I accidentally put some of them in backwards and got a higher voltage reading. I re-tested them and they all stayed the same values as before. Could they be damaged by doing that?