DrFrankencopter

Re: 1176LN Rev D DIY
« Reply #1240 on: January 27, 2010, 11:21:57 AM »
Also, is there a schematic for Mnats 1176D?

Yes there is: http://mnats.net/1176_revision_d.html

-Paavo

Many thanks....

A note for Mnats, the link to the 1176 rev d page from this page: http://mnats.net/1176.html does not work (it needs .html added to the end).

Cheers

Kris


DrFrankencopter

Re: 1176LN Rev D DIY
« Reply #1241 on: January 27, 2010, 11:30:28 AM »
I checked the schematic for the Mnats RevD, and there is no T&C resistor from the 20:1 switch to the sidechain input. This resistor is present on the Urei Rev D schematic, and on the Purple MC76. It looks like it will affect (raise) the threshold for the 4:1, 8:1, and 12:1 settings.

Any reason why this resistor isn't included in the mnats design?

Cheers

Kris

trancedental

Re: 1176LN Rev D DIY
« Reply #1242 on: January 27, 2010, 11:54:38 AM »
I have a new guide that explains it better.

http://hairballaudio.com/docs/FET_Compressor_PCB_Board_Info_V2.pdf

The GRN connects to the bypass switch.  When pad 22 is connected to GRN on the ratio board you are compressing.  On the earlier documentation I indicated "Attack Pot" because I have always put the GR bypass on the attack pot.  Wrong on my part.

Check out the new doc.

Thanks,

Mike

Thanks Mike, Thats all sorted out now. Just one more silly question, which is the exact positive / negative + & - connections for the VU meter on the Hairball Pushbutton Meter PCB?

My 1176 is passing audio & seems to be compressing but the meter is not responding so far, so I just need to confirm the meter wiring as I've tried it both ways without much difference ???

dmp

Re: 1176LN Rev D DIY
« Reply #1243 on: January 27, 2010, 02:04:11 PM »
I think the top VU is + on the Hairball Meter PCB... the pads are: Y  X  28  29   VU(-)  VU  (+)
That's how I hooked mine up and the meter is working correctly for signal and GR.
Here is a wiring diagram (based on the really nice one Skylar made) that uses the Hairball supplies... like the EA-5002, pushbutton switches with the hairball pcbs, the hairball attenuator, the attack SPDT switch from Mouser, and a zero set pot on the hairball faceplate.
Sorry this is messier than Skylar's nice original ;)
 

mnats

Re: 1176LN Rev D DIY
« Reply #1244 on: January 27, 2010, 03:16:32 PM »
- In the calibration video, mnats says the Qbias trimpot should be adjusted all the way to ground, and then dialed back to bring the fet into conduction. I think this is reversed actually, the Qbias should actually be adjusted fully negative, since the fet Vsource is equal to ground. The fet becomes more conductive as Vgs approaches zero, I believe, i.e. when Vg is more negative the the resistance of the fet is high, i.e. no gain reduction of the input signal.

I think this was also mentioned by the uber-helpful gswan in the "All thing G1176" thread. I will correct the video one day, but I guess most people will end up turning the trimmer without getting out a multimeter or looking at the spec sheet anyway!

Any reason why this resistor isn't included in the mnats design?

Or the G1176? The prevailing wisdom when the board was first made was that the T&C part was done to confuse cloners. Even if not true, it worked!

If you are motivated to go through the measurments, it's easy enough to add, isn't it?

DrFrankencopter

Re: 1176LN Rev D DIY
« Reply #1245 on: January 28, 2010, 09:06:36 AM »

Or the G1176? The prevailing wisdom when the board was first made was that the T&C part was done to confuse cloners. Even if not true, it worked!

If you are motivated to go through the measurments, it's easy enough to add, isn't it?

Sure, it's no problem. I only even noticed it because I had 'reverse engineer' the Hairball Audio ratio & meter boards against my Purple MC76 schematic, and found that the T&C resistor was missing (and the 10 Meg resistor was in a different position). I was having some issues with my 20:1 ratio, and thought that the lack of T&C may have been the root cause. In the end, my problems were a combination of a 5.6k resistor instead of 56k, and a poor solder connection on one of the 20:1 switch poles. I'm going  to do some subjective listening tests and see if the thresholds are good enough without the T&C.

Switching subjects...On my MC76 when doing the GR meter tracking adjustment, I ended up needing 2.5 kOhms resistance for R44 to make the meter zero, and track 10 dB of gain reduction. I notices on the Mnats 1176D, R44 has a maximum value of 1k. Is this indicative of a problem with my build (I didn't use matched FETs)? My meter read less than 10 dB GR when R44 was set to1k, and the only way to bring it up was to increase R44. Perhaps a 2k pot for R44 might allow for more variability...

Cheers

Kris

edanderson

Re: 1176LN Rev D DIY
« Reply #1246 on: January 28, 2010, 11:16:05 AM »
on the mc77 we used a 2k trimmer for r44.  it all depends on the FETs -- some will bias with just a 1k trimmer, most with a 2k, but i've seen some people use even a 5k trimmer on their mnats builds.  the FETs are really all over the place and if you aren't doing some selection beforehand, 1k may not cut it.  i think the screen on the v2 mnats PCB even indicates a 2k trimmer for r44, but i could be wrong about that.

glad to hear that you're getting closer kris.

ed

DrFrankencopter

Re: 1176LN Rev D DIY
« Reply #1247 on: January 29, 2010, 09:18:58 AM »

glad to hear that you're getting closer kris.

ed

I'm 99% done Ed...many thanks for your help and encouragement. The only thing left is to do the distortion calibration, and decrease the lamp current. I used a 7812 fed from the 40V post recitifer to power the lamp, and despite being on a decent heatsink it's getting too hot for my tastes. I'm going to experiment with a series resitor with the lamp and see if I can get the current down while still having some brightness. If that doesn't work, I'll try subbing out the lamp for some LEDs.

Cheers

Kris

edanderson

Re: 1176LN Rev D DIY
« Reply #1248 on: January 29, 2010, 10:46:11 AM »
since the 7812 has to drop the voltage from ~40VDC down to 12VDC, that will be a significant amount of power unless you reduce the current in the lamp quite a bit.  you could put a resistor in series before the 7812 to drop the pre-reg voltage, and keep the current in the lamp the same.

if you have a way to look at the THD signal on a scope or do an FFT of the signal, you may be able to fine tune the distortion measurement.  there is a range on the trimmer that will give you a minimum value, but within that range you can adjust the harmonic content a little if you have the means to measure it.

ed

DrFrankencopter

Re: 1176LN Rev D DIY
« Reply #1249 on: January 29, 2010, 12:26:55 PM »
since the 7812 has to drop the voltage from ~40VDC down to 12VDC, that will be a significant amount of power unless you reduce the current in the lamp quite a bit.  you could put a resistor in series before the 7812 to drop the pre-reg voltage, and keep the current in the lamp the same.

if you have a way to look at the THD signal on a scope or do an FFT of the signal, you may be able to fine tune the distortion measurement.  there is a range on the trimmer that will give you a minimum value, but within that range you can adjust the harmonic content a little if you have the means to measure it.

ed

I've already got a 75 ohm 3W resistor before the 7812, which should drop the voltage by 7V. But there's still something like 2W being dissipated in the heatsink. I'm going to try looking for alternate lamps...perhaps I can skip the 7812 entirely.

For the THd measurement I was planning on using Wavelab's FFT analyzer, and a 1kHz sine input...hopefully it'll be good enough to show gross adjustment problems. I'll attempt to adjust for minimum high harmonics.


Cheers

Kris


Biasrocks

Re: 1176LN Rev D DIY
« Reply #1250 on: January 29, 2010, 12:47:30 PM »
Do you need to run the lamp DC?

On my build I ran a 220 or 240 ohm resistor off the power transformer secondary to nicely light my Hairball meter.

Mark
http://SharktankPro.com

"I'd rather use an SPX90 than a UA plugin....." Joe Barresi

trancedental

Re: 1176LN Rev D DIY
« Reply #1251 on: January 29, 2010, 08:04:17 PM »

Here is a wiring diagram (based on the really nice one Skylar made) that uses the Hairball supplies... like the EA-5002, pushbutton switches with the hairball pcbs, the hairball attenuator, the attack SPDT switch from Mouser,
 

Cheers Dmp ;D

On your wiring diagram, there's two wire connections leading from the Gain Reduction Bypass switch function included on your DSPT switch (Attack Pot) used on your build. From the off position lug one goes to Ratio board BLK & one to PCB ground (on the Mnats board "G" I assume?)

I have wired my Gain Reduction Bypass / disable function on the meter pushbutton board (instead of the usual on / off power switch)  as indicated  in Hairballs instructions http://hairballaudio.com/docs/FET_Compressor_PCB_Board_Info_V2.pdf the wiring layout I used is the same as in the last diagram instructions with the 22 going to the main board & the BLK & GRN going to the ratio board.

So I don't actually have the extra wire fitted which in your case goes to the point "G" on the main board, I have checked for continuity between the "G" & "22" when I engage my GR bypass switch without success.

So do I need another extra wire from the BLK lug on my switch to "G" then?

Cheers JK


Hairball Audio

Re: 1176LN Rev D DIY
« Reply #1252 on: January 30, 2010, 02:50:06 AM »
Ratio BLK and Bypass BLK need to be connect together as well as to signal ground at some point.

Usually the shield from 22 carries the signal ground from the board to the bypass or ratio switch.

Mike
Hairball Audio, LLC
DIY Parts and Kits
http://hairballaudio.com/shop

Top secret stuff and upcoming releases:
http://facebook.com/hairballaudio
http://twitter.com/HairballAudio/

dmp

Re: 1176LN Rev D DIY
« Reply #1253 on: January 30, 2010, 11:30:41 AM »
JK, I think it is a little confusing because in the sketch the connection to signal ground for the ratio and bypass BLK is connected through the wire shielding, as mike says. This is an unusual way to do things - usually a shielding is only connected at one end and a seperate wire would be used to connect signal ground. But, it is easier and works fine in my build.
You should have continuity between PCB "G" and Ratio "Blk". If you do not, you can run another wire to connect the signal ground between the two.

patterson

Re: 1176LN Rev D DIY
« Reply #1254 on: January 30, 2010, 02:17:02 PM »
I got the kit from Hairball and the revD worked right away. Having only used software versions of the 1176 before, I am amazed at the difference it makes to use the actual unit.  Very musical compression. Thanks to you and to Mnats for a great project.

I noticed that some people are using tantalum caps instead of the cans in some cases.  Does this make a difference in the sound?

edanderson

Re: 1176LN Rev D DIY
« Reply #1255 on: January 30, 2010, 02:56:09 PM »
in some of the older original units, two caps in the sidechain were tantalum because they were lower leakage than the electrolytics available at the time.  most electrolytics today aren't leaky enough to cause any issues, so it isn't a big deal.  but you can safely use tantalums if you want. 

i think some people might notice a sonic difference if you use a tantalum cap in the audio signal path, but just as electrolytics have changed, many of today's tantalums are different from the old ones (for example those found in old neves).  feel free to experiment with different caps once you've got it working.

ed

trancedental

Re: 1176LN Rev D DIY
« Reply #1256 on: January 30, 2010, 07:53:40 PM »
Cheers Echo North & Smd, I'll have to go through the exact wiring & check everything again, bit arkward as my home internet isn't working at present so troubleshooting is now harder ::)

Anyway I had the 1176D on for 5 hours today, the board voltages are -9.6 / + 29.8 which is hopefully just within range & the heatsink seems cool enough.  I ran tracks through it & tried some DI bass & guitars. Sounded very nice with my '67 Gretsch Country Gent ;D

Its seems to be working & compressing but with some slight problems, VU not working (I must have this wired up wrong  ???  ) but I managed to do the q-bias & some of the other calibration tasks by using another external VU meter.

Attack switch full clockwise starts to stutter especially on drum tracks, & the -20 & -4 compression settings have a very low level bubble type noise, the other compression settings are fine.

Just a matter of doing some more checking & rewiring should finish it off


T-rex

Re: 1176LN Rev D DIY
« Reply #1257 on: February 04, 2010, 09:35:00 PM »
Is there an audible difference between using the 2N5088 and the 2N3037s?  I have a set of both and was going to maybe socket the board this time so I could switch between them but before I do that is there enough of a difference to matter?  I tried searching but as you can imagine "transistor" and "sound" brings up pretty much every thread. . .

dmp

Re: 1176LN Rev D DIY
« Reply #1258 on: February 10, 2010, 04:33:26 PM »
I've been looking at the signal preamp section of this cicuit for a couple days and I'm wondering if anyone has figured out the design? I'm having trouble as the gains I'm calculating seem really large. Total voltage gain of ~100 dB? Does anyone know if this is expected? Voltage gain of Q2 >2000!  I don't have a circuit simulation program so I may have made a mistake in the equations. Basically, I used ohms law and transistor rules to simulate the preamp in an equation solver. 
I started looking at it because a fairly large input signal (a few volts pk-pk sine wave) shows negative clipping due to (I think) the bias of the emitter follower, Q14, but also perhaps by the high gain of the preceding amp stage, Q3. I thought maybe changing the bias of Q14 might be an improvement in headroom since it is at ~4v - but I first wanted to understand why the revD had R12=920ohms (later revs increased this I think, maybe to increase the headroom at this stage?).  But I'm struggling to understand how the distortion trimmer changes the frequency response and if this feedback to Q2 is sensitive to R12, which sets the bias back to Q2.
I guess this is a build thread - if there is a better thread for this question I apologize
Dan

Re: 1176LN Rev D DIY
« Reply #1259 on: February 10, 2010, 06:40:55 PM »
Is there an audible difference between using the 2N5088 and the 2N3037s?  I have a set of both and was going to maybe socket the board this time so I could switch between them but before I do that is there enough of a difference to matter?  I tried searching but as you can imagine "transistor" and "sound" brings up pretty much every thread. . .

Socketing them all is kind of nice, anyways, and since you already have the two different types . . . go for it!    I'm sure alot of folks would be interested to know your results.