Official Forssell Limiter Build thread

[Nele]

Farnell sells a reverse log 5k Alps pot:

farnell-link

number 919-524

adios,
corneel

 

[ChuckD]

Tommy.

NEWARK has those REV log 5k pots coming in now and they are due to ship out on the 3rd of DEC.

FYI

-ChuckD

 

[tommytones]

Cool! Thanks.

I'll dig up the part numbers in a bit but it's easy to find.

 

[andre tchmil]

Hi Nele,
does Farnell also stock 20k or 25k or 10k rev log ?
can't find any .

 

[tommytones]

Finally got my parts from Allied :roll: Man, Digikey in Canada is the best! I've ordered and received stuff 3 times before I got my stuff from Allied...

500K lin 308N-PC
5K rev log 91A1A-B24-T13
5K log 91A1A-B24-D13
100K log 91A1A-B24-D20

Quick note on the parts for the pc boards. The BC caps listed for the 1000uF/10V (C37) are not going top fit properly. I've ordered both part numbers and it seems, for whatever reason, BC is making these in a larger size than they did before :roll: There's enough room to put it, but you'll have to bend 1 of the legs a bit more than usual. This is only relavent to the 5534 output driver section. If you're going discrete, then no worries.

Anybody put one together yet?? I'm gonna finish mine off tomorrow and fire it up.

 

[MikoKensington]

OOOHHH. Fire at will! I'm excited.

 

[tommytones]

Geeze, I was really not in hurry to put this together and by the sounds of it, i thought I was gonna be the last one to do so. Looks like I might be the 1st...

 

[Kev]

ToM !

don't forget
PICTURES

I want some pictures ... anything will do

 

[tommytones]

Okay, it's snowing, I'm unhappy (grave understatement), and it's cold....time for me to gripe !

I've fired up the comp today and have a few comments off the bat.

1st, I've found that (surprise) I had to play with the psu in order to acheive symetric output. I'm not sure if it's a result of using 317/337 instead of the devices that Fred had in his schem, but the resistor values don't work out. It's close, about 25.6v but I'm picky about this. What I've done is replaced the 6k19 with 4k75 and parallelled the 300r with a 500r pot and a 120r in series (didn't have a smaller value pot so I took the long road). Result is a trim range from about 19-27v. and the ability to really dial in the voltage wanted.

The following is in reference to the non discrete build.

2nd, I measured an unwanted roll off begining @ ~500Hz. It's quite steep too. With or without a transformer. I didn't play around with it very long but looking at the schem, maybe we can find an explanation of what's going on.

Schem

3rd, the way that butta has the pot going in that schem, it acts as a fader and not a make-up adjustment. Again, I think it might be best to tie a pot in the feedback to vary the gain of this stage. I didn't try it out yet but I don't see why it wouldn't work.

Now, I must say that I didn't have the sidechain active at the time these tests were made. It was bypassed by simply not installing the bypass switch which feeds the sidechain. I don't see how would cause any issues but I may be wrong.

Comments people.........

 

[ChuckD]

OK Tom

I am getting close to finishing mine a well. A few questions:

1. What transformer spec did you use? Did that cause the difference in your design? I was going to use the +/- 22V Torriod from Digikey.

2. C43 what is this guy's value?

3. Which U9 gets the OPA2604 when hooking up stereo units.

4. LED's? how do these get installed and what kinds can I use with this chip?

I am still waiting for Fred's Op-Amps to arrive but then I should be good to complete it.

-ChuckD

 

[ChuckD]

Also Just noticed under the Big cap I was missing another:

5. C4 ?? What is this guys' value?


Lets see some pictures of the finished unit!

-ChuckD

 

[Infernal_Death]

I already asked Tom about C43 and he said it's ok to use a 33uf cap. I used the same one as for C41,C42,C7 and C8.
As for the other questions i also would like to see an answer.

My 4 boards are stuffed around 90%. Just some caps missing, the discrete opamps for 2 boards and i still need to buy the transformers and the revlog pots.
Was it only me or did anybody also had some problems fitting the 1N4004 diodes and the connectors ? I couldn't get the diodes through the holes so i soldered them on the top side.

Flo

 

[ChuckD]

haha !

Yes those holes were way small eh?

I actually soldered new leads on all of them from clipings off the resistors! Too funny I thought I ordered wrong then I ordered again and got the same size diodes the second time.

Thanks on that missing part number, bad news is I used my last of the 33pf caps already now I have to order again

-ChuckD

 

[tommytones]

For stereo, both U9 and the 1K (r45) following are stuffed on both boards.

Led's, any typical ones can be used or a 10 segment Led blocks can be used. Install according to the schem.

I was feeding 29v dc to the on-board psu. I have the same 22v toroid from Digikey and plan to use it when I put these in a box. I did the trimmer thing because I like to feed exactly the same voltage to both +/- rails. You can use the indicated values with a bit of offset between the two. As is, I'm feeding +/- 22v.

Sorry about the diodes. It seems that some boards were okay and others were too small. The through plating was a bit thicker than normal this time round. Actually, a lot thicker. I used the same size holes as I always do.

It's a compressor!!

I fully stuffed 1 board yesterday and it works quite well. Don't ask what it sounds like cause I haven't passed anything other than tones through it. R28, I'm using 100r. Works well for me.

There are some mods that need to be made to the 5534 output driver. Unity gain adjustment was the 1st thing to fix. As is, there's about 8db gain in to out. I also had to add a 10K resistor from RV1 leg 1 to leg 3. This replaces the pot in that position and maintains stability. I also added a series resistor from RV1 leg 1 to Leg 2/wiper. This will set the gain down to about 4 db gain. If you tie another 10K across the wiper and RV1 leg 3, this will bring the gain to about -2db in to out. This is where I think I'll settle for my units. I'm gonna draw this up later today so you can all see what I'm talking about.

Next is the make-up gain. I found that placing a 100k pot in the feedback (keeping the 10K resistor in series) will work out good for gain. This gives about 24db of clean gain which is more than ample I think for this unit. Again, I'll draw it out for you.

Now, the biggest problem with this output circuit is that there's severe roll-off starting @~500Hz. I'm measuring 10db down @ 47Hz...not too good. I've tried several things without success. Replaced c39 with a larger 1000uf, no change. Bypassed the 100r series (r39), no change. I'm a little stuck on this one. Referring to the schem I linked above, what do you think is happening.

I know it's the output driver because I measured at the input of it and it's fine.

 

[ChuckD]

Tom,

Thanks for those values and changes.

The link is dead? I think?
Can anyone else see it?


-ChuckD

 

[DrFrankencopter]

Now, the biggest problem with this output circuit is that there's severe roll-off starting @~500Hz. I'm measuring 10db down @ 47Hz...not too good. I've tried several things without success. Replaced c39 with a larger 1000uf, no change. Bypassed the 100r series (r39), no change. I'm a little stuck on this one. Referring to the schem I linked above, what do you think is happening

Perhaps try a larger decoupling cap leading to ground (C37 going by the link above). My guess is that too much AC is being shunted to ground around the feedback path. Maybe increasing the value of the 2.2k resistor would help too.

Cheers,

Kris (just starting on this comp myself)

 

[tommytones]

Sounds like a plan Kris. I was wondering the same thing because I remember there being some talk about a 4700uF originally. I'm gonna try out a few things tomorrow along with this.

Stay tuned for the drawings...

 

[tommytones]

Here is the schem with the changes

It's only the section that I changed...there's still the stuff before and after what is shown.

 

[rafafredd]

why don´t you do R38 variable instead?

 

[tommytones]

Well, r38 needs to be there even if the pot is placed at this point, no? I wanted to show a visual of what I did. The pot could be placed in series with r38.