Thanks a lot Gustav, really sorry about this misunderstanding. This unit is really great ! During the cakibration it was in dual mono, but 2n5457 are not matched on my unit...i think about taking 4x2n5457 matched on hairball (fairchild old series). On the datasheet it seems that i have to put those transistors reversed ( DSG to GSD), am i right ?
Build Thread:MS76
- Thread starter Gustav
- Start date
Help Support GroupDIY Audio Forum:
hugo said:
I am not sure whats giving you the matching problems, but if you feel trying the Hairball transistors will work better for your build, you can give it a try.
The board is designed for the 2N5457. I am not aware that different pin-outs for that part exist (see attached for reference), but if you have become aware of a difference in the part you intend to use, you should adjust it, of course..
Gustav
Attachments
rprimmett said:
Hi Ryan.
Blend should work as a "crossfader" between the clean and compressed signal, so one end 100% compressed, other end 100% clean. Which end is what will depend on the orientation of your wiring on the pot.
It could be a number of things, including some weird combinations of what you are feeding the unit, and what your settings are.
1. Make sure you are in dual mode (in MS mode, if you only send signal into the M channel, you will have no "clean" signal on the S channel blend knob, but depending on your settings, you can have signal on the compressed end of the S channels blend knob (check this signal flow).
https://pcbgrinder.com/image/cache/catalog/FLOW/MS76FLOW-900x900.gif
In short - make sure you are not in MS mode, and check one channel at a time.
2. Starting with the left channel in split mode, check for signal on the points where the clean and compressed signals are tapped from, then trace them to the pot.
3. In single channel mode, you will always have the clean signal on the blend pot, but the compressed signal will only be there if you are actually sending it through the unit (in/output pot can cut it off).
Thats should get you started.
Gustav
rprimmett said:
Test points are spread through the signal path. They are marked, but not named appropriately in the schematic.
What you should see depends on what you put in, and they are just handy for poking around with a scope.
Gustav
frederickalonso
Active member
- Joined
- Jun 27, 2012
- Messages
- 40
Also big thanks to Gustav, great documentation!
Everything works fine, soft compression dynamics like I like them.
In M/S mode my mono is not complete in the center, so I will try to insert Lundahl transformers to solve this.
Hope it works..
Video:
https://www.instagram.com/p/BJSwyLeDlLs/?taken-by=alonso.frederick
Everything works fine, soft compression dynamics like I like them.
In M/S mode my mono is not complete in the center, so I will try to insert Lundahl transformers to solve this.
Hope it works..
Video:
https://www.instagram.com/p/BJSwyLeDlLs/?taken-by=alonso.frederick
frederickalonso said:Also big thanks to Gustav, great documentation!
Everything works fine, soft compression dynamics like I like them.
In M/S mode my mono is not complete in the center, so I will try to insert Lundahl transformers to solve this.
Hope it works..
Video:
https://www.instagram.com/p/BJSwyLeDlLs/?taken-by=alonso.frederick
Well done, Frederick!
If you dont mind, can you tell me how you are measuring that mono is not completely down center in MS, have you traced it back, and how much is it?
In theory, anything that is M content will per definition be center, anything that is not will per definition be S, even if they are not in complete balance as far as what you are putting into the unit.
I usually have to have a unit on the bench to double check, but I dont see how mid shift could happen anywhere but after the decoding, and I dont see how it could be fixed by swapping the IC outputs to transformers.
Hope to understand this, and see if we can get you 100% there!
Gustav
frederickalonso
Active member
- Joined
- Jun 27, 2012
- Messages
- 40
Gustav said:frederickalonso said:Also big thanks to Gustav, great documentation!
Everything works fine, soft compression dynamics like I like them.
In M/S mode my mono is not complete in the center, so I will try to insert Lundahl transformers to solve this.
Hope it works..
Video:
https://www.instagram.com/p/BJSwyLeDlLs/?taken-by=alonso.frederick
Well done, Frederick!
If you dont mind, can you tell me how you are measuring that mono is not completely down center in MS, have you traced it back, and how much is it?
In theory, anything that is M content will per definition be center, anything that is not will per definition be S, even if they are not in complete balance as far as what you are putting into the unit.
I usually have to have a unit on the bench to double check, but I dont see how mid shift could happen anywhere but after the decoding, and I dont see how it could be fixed by swapping the IC outputs to transformers.
Hope to understand this, and see if we can get you 100% there!
Gustav
Gustav,
I use my Waves ( Analyser vst ) on my input channels in Ableton Live. On my other gear everything is fine. If I switch the unit to dual mono everything is fine = playing a mono kick thru it and it comes back in center at both channels. If I play a full stereo track all is fine. Only M/S is a little off, don't know why. If I adjust the L/M gain (on PCB) I can fix this issue but once I switch to dual mono and play a full stereo track the bass is more to the left. In M/S mode I tested to mute the volume from the Side (R/S) so I could see what's going on and then I see that it's not complete center. Hope you understand what I try to write here ;-)
So in perfect stereo (dual mono) all is fine. In M/S the mono is like around 11 hour (more to the left) instead of 12 on a clock where 12 would be the center (mono)
PS: I tried to switch the THAT's from Right to Left but all is the same.
If I know where the calculation M/S is done I can maybe adjust something or add an extra resistor maybe?
Well I have seen some articles where it's obvious that Lundahl transformers can give a perfect focus and avoid stereo artifacts. I have a pultec EQ for example with the same Lundahl transformers and the difference I had by replacing FET outputs in this EQ with the transfo's was huge. That's why I think it will help as well.
BTW: I have had an API2500 and Elysia xpressor 500 and price/sound wise I would rate this unit between 1400€ and above the Elysia. Seriously, I like it. Great build as well.
frederickalonso said:
So happy you are into it
your drift of center is obviously not due to the output stage - and I can't even wrap my head around a possible error to cause it, other than something funky going on in the MS decoding. Adding the Lundahl output trafos will certainly not have any effect on it.
Please mail me, I may have an idea to check, but Ill need to add some documents, which is not so easy here.
Gustav
frederickalonso
Active member
- Joined
- Jun 27, 2012
- Messages
- 40
pstcho said:Hi guys, question for the people having finished the unit
HOW DOES IT SOUND
Is it close to a classic 1176, cleaner, muddier, more/less distortion ?
( i guess i will go for a unit pretty soon )
Thanks
For me it sounds open and has a little but warm saturation to it.
On a complete mix you can have more click (attack) on drums or give
the bass more body. I also found out that a resistor change on the release (PCB) can give more tail to drums.
I'm thinking to build a GSSL and MS HOL from pcbgrinder as well. Not yet sure wich one will be my "sound". If someone has examples, please share.
Just finished wiring this guy up. I'm having a few issues that I can't wrap my head around.
1st Meter Level seems low. I changed the 20k resistor in for the meters back to 10K and I was able to get it to 0.
Input pot level does not affect what the Meter is reading. With the Input all the way to 0 ( counter clockwise) the meter is still reading 0 VU. Looking at the Test Point 2 I see the signal level change with the pot. Same for the Output Pot with the meter set to read the output. Output at the xlr does change.
Any thoughts? I'm assuming its something with the meter switching but I can't find anything wrong.
Thanks for any help!
1st Meter Level seems low. I changed the 20k resistor in for the meters back to 10K and I was able to get it to 0.
Input pot level does not affect what the Meter is reading. With the Input all the way to 0 ( counter clockwise) the meter is still reading 0 VU. Looking at the Test Point 2 I see the signal level change with the pot. Same for the Output Pot with the meter set to read the output. Output at the xlr does change.
Any thoughts? I'm assuming its something with the meter switching but I can't find anything wrong.
Thanks for any help!
RSRecords said:
Not sure what you mean by "able to get it back to 0", or which resistor exactly. Could you describe the problem leading you to make the change, the reasoning for making it, and specify which 0 (0 in GR mode, 0 resting in VU mode)
RSRecords said:
I would start by tracing the signal to the meter, and also make sure the switching is actually done on the relays (check the voltage going there)
Gustav
Gustav said:
By 0, I meant 0vu while in VU IN mode.
I was referring to the 20k resistor in the feedback loop of the tlo72. In the "Errata" section you say to replace it with a 10k for proper gain. I was attempting to complete the calibration but the 1st step is to set the level to zero vu in which I was unable to do. I've changed the resistor back to 10k since it did not seem to be the issue.
Gustav said:
Voltage is at the relays and I can hear them switching. Everything looks good and I went through and reflowed the solder on the relays for good measure. I measure 12v at the relays depending on the meter selection switch.
Are test points 2 and 3 after the in and out pots respectively?
I should also note that both channels are behaving this way.
