Gates style build - HF roll-off

[gg85]

Wondering if anyone might be able to help me with some HF roll-off in a vari-mu build. Over the past 3-4 years I’ve been slowly building and modding a dual Sta-level style build. It’s at the point where several things are quite different from the original, including the metering and output section (post transformer), so I wouldn’t exactly call it a Sta-level, although the audio path is the same as the original. I’m using the Lundahl 1922 on the input and Hammond 1650G on the output. I’ve attached schematics of the output section design.

Basically, it’s an active output with a 600-ohm resistor strapped across the secondaries. There’s a switchable 10dB H-Pad via a relay. It switches between either the 600ohm load resistor or the h-pad. This is then unbalanced via a THAT1246, fed into an OPA1656 with a rotary in the feedback for 0.5dB steps. Then it’s balanced again via THAT1646 and then XLR out. My goal was to turn it into a much more subtly controllable compressor, with .5db steps on ins/outs.

There are a few reasons why I am using this output design, but the main one was an attempt to fix a frequency roll-off issue. Previously I was doing it passively with a 600ohm bridged T on a 2-pole rotary switch. I noticed when I was at unity, the response would flatten out more (but still a little HF roll-off), and at more attenuation via the switch, it would roll off more from about 8kHz. I thought maybe having a constant 600ohm load with an active attenuator might be another way to do it.

With it all hooked up, on the scope I am seeing a completely flat response at the secondaries, after the load resistor, and at the input of the THAT1246. After the THAT1246 I am seeing the roll-off. The amount of roll-off does not change at different levels of attenuation, so I am happy about that, but I can’t figure out why it’s there at all.

In the THAT1246 datasheet, I can see each input of the opamp has a 12k resistor, which I thought should set the impedance high enough to work.

Feeling a little out of my depths at this point. Any help would be amazing!

 

[radardoug]

So its flat coming out of the transformer? Then why are you bothering with the SS stage? Just get a 600 ohm attenuator or switch to do your .5 dB steps. Your process shows a high degree of anal, music is not like that.

 

[gg85]

So its flat coming out of the transformer? Then why are you bothering with the SS stage? Just get a 600 ohm attenuator or switch to do your .5 dB steps. Your process shows a high degree of anal, music is not like that.

I had that originally but could hear the roll-off, which led me to get the scope out.

My goal has been to use the compressor for mastering. I agree music is not like that, but sometimes mastering is, especially when you want to recall. 0.5dB steps come in handy.

The SS stage has been an experiment because my previous attenuator gave me a roll-off too. I used the attached design, which I got somewhere off this forum, it was a while ago so I can't remember who posted it.

 

[radardoug]

Ah, so you did have a rolloff straight off the transformer? Then you need to deal with that first. And the bridged T is not a great attenuator, due to the changes in loading.

 

[gg85]

It's been on the back burner for a year so I can't remember if it was flat at the transformer when using the bridged-t, but I know for sure with the current active output it is flat after the transformer, flat after the 600 ohm load resistor, and not flat after the THAT1246, which confuses me.

I wasn't aware that bridged T's change the loading at different positions. What would you suggest as a better attenuator than the bridged T if I were to go back to passive out?

appreciate your help!

 

[radardoug]

H pad, but you need more switch poles.

 

[gg85]

Just swapped out the 600r load resistor for 1200r and my response is flat from 20Hz-30kHz.

I thought 600 would be correct as that's what drip's build was asking for in a variable t-pad attenuator. Not sure if there's difference in loading with a 600 ohm variable t-pad vs 600 ohm load resistor. I get flat results with the bourns 600 ohm variable t-pad, but a roll off with the 600 ohm load resistor and no attenuator. Change it to 1200 ohm and it's flat

 

[radardoug]

Are you terminating the Tpad? Thats how they were originally done. But the answer is probably that the bourns Tpad ends up being 1200 ohms.
Now you know, you could build a 1200 ohm Hpad.