What's the idea behind this?

[Viitalahde]

http://www.manleylabs.com/containerpages/de-ess.html

I think we discussed it a little at the old place. I also tried to get the schematic from Manley, no luck but EAM answered some questions.

If I've understood it right, you have a series LC filter, and feeding it with DC makes its inductance vary. Or? :?

 

[ruffrecords]

It does not alter the value of the inductance but its quality factor Q presumably by altering the value of a series or parallel resistance. This simply changes the width of the resonance.

HTH

Ian

 

[burdij]

This appears to be an application of a magnetic amplifier, a form of saturable reactor. Here is a wikipedia article. There are lots of other references.

http://en.wikipedia.org/wiki/Magnetic_amplifier

 

[ioaudio]

"A typical simple mag amp contains two identical coils, each having identical high permeability square loop magnetic cores and each wound with an identical winding not shared with the other coil."
from: http://www.butlerwinding.com/elelectronic-transformer/mag-amp.html

vca?

 

[Sorr]

"The Manley High Frequency Limiter was originally based on the same operating principle as found in the old Ortofon lathe cutting electronics de-Esser"

Anyone have the schematic for this?

 

[Viitalahde]

OK, I think I'm getting the principle. Saturating the core makes the inductance drop and thus it "opens" a notch at the selected frequency..?

But how does the impedance of the coil in a saturable reactor act?

If I am getting this right at all, having an LC series filter to ground would require the coil to have a high Z when not saturated, thus making the filter ineffective. When approaching saturation, Z would drop and the notch would go from wide/shallow to narrow/deep.

The Ortofon de-esser schematic would probably give some light to this..

 

[bcarso]

Changing L changes f, not Q.

 

[JohnRoberts]

[quote author="mediatechnology"]

The MANLEY HIGH FREQUENCY LIMITER works with entirely passive 3-component front-end circuitry sample-steered by a solid-state side-chain which energizes a "variable reactor" coil varying its "Q". The now-controlled programme is then amplified by an all-tube line amp with absolute sonic cleanliness. Selectable frequency ranges: 7 - 16KHz & 5 - 9KHz.

Brad: Do you suppose that they're saying one thing and meaning another? The quote seems internally contradictory.[/quote]

My interpretation is they are using a "saturable reactor" a specialized coil or coupled coils where they can manipulate the impedance of one winding by varying the flux density in another on the same core. These are, or were common in power supplies for current limiting and such.

So they appear to be using a solid state side chain to modulate this magic coil in their all tube tweako audio path. Bigger question for me is why does this sound attractive, but I guess I understand the attraction to idiosyncratic technology.

JR

 

[bcarso]

[quote author="mediatechnology"]...But we rule out VCAs and Fets by virtue of all tube.[/quote]

With sufficient heroics/madness you can make a pretty accurate VCA with tubes, but I doubt that is what they are doing.

Reminds me of a dual photomultiplier preamp and analog divider box I made for someone at UCLA once. It was a substitute for a large rack-mount unit at Mt. Wilson/Hale Observatories, which had a quarter-square multiplier at its heart, in turn using a specially selected tube with a square-law characteristic over a limited but evidently useful range.

The function of the units was for photographic densitometry, essentially dividing out the photographic plate sensitivity from the spectrum adjacent to it.

 

[Crusty2]

What about faking the reactor dealy thing like this? (Replace pot with FET). I think LDR would be too slow, it'd "wah".

(Or am I on Cordazine again?...)

 

[thelivingroom]

I’m digging up this ancient thread because I actually have one of these, and started looking into repairing it this week for a new challenge and to maybe sell to help fund finishing up too many projects I’ve started from in here. Pretty much nothing comes up online, but of course this site popped up haha. It was given to me about 15 years ago by a studio owner here who had moved onto mostly plugins, along with a nice little pile of other gear. This piece went straight to my storage room, because when trying it out it would crackle on the initial attack after it crossed the threshold and didn’t sound like it was actually doing any reduction. They only made 36 of them, Manley doesn’t have the full schematic, and in their tech notes the 3 they’ve repaired in the past 28 years didn’t have the same problem as mine. The partial schematic I received is for the PSU and line amp, which are both perfectly fine, so I’m gonna have to map out the sidechain section and draw a schematic to try to figure it out.

 

[thelivingroom]

I dove into this, mapped out the side chain PCB, drew the schematic for it, and read about saturable reactors last weekend before starting to troubleshoot. It had two issues, the main one being that I noticed R33 (top right in the schematic) was for some strange reason connected to pin 13 and its associated trace, therefore bypassed by the trace and doing nothing. In the one somewhat clear picture I could find online, from a Reverb sale 7 months ago, I saw that it should be in the feedback loop between pins 13 and 14. I fixed that and the DC feed to the reactor now seemed correct when working the threshold (from 11ish volts full CCW down to 0 VDC at full CW), but it still had the crackle. I scoped every output pin on the TL084's and when I got to pin 14 that feeds the reactor, sure enough there was AC showing up in that line, probably from the PSU/line amp PCB that's directly behind the reactor. I increased cap values from that line to ground until it wasn't showing up on the scope anymore, and now it's clean and everything appears to work properly up to the reactor, but I'm still getting no reduction on output when testing it with a drum overhead track, vocals, or signal generator. I've tried different combinations of adjustments to RV2 & 3, numerous frequency ranges from 5k to 18k, sweeping all the settings, etc. My one fear is that maybe the reactor is toast, because I have found nothing online about it. Searched the TDK model number and a ton of combinations of other info that's printed on it and names that it may be called, and nothing. The resistance on the AC side of the reactor is measuring really low, like 2.2 Ohms, and I don't know if that's how it should be or not.

Maybe @gyraf has some insight?

 

[gyraf]

Upper L corner, 100uF:

"Added this to filter out AC
that was getting in this DC
line and wreaking havoc
"
..you don't want this extra electrolytic here, it messes seriously with your time constants and upsets the current-drive opamps. If there's excessive AC component here, it have to come from the sidechain or driver

Current through reactor/attenuator is differentially driven between 1) U3A, TL084 Pin1, adjusted through RV#3 and 2) driven from inverting amplifier U3D, pin14

this also drives panel meter

Input signal for all this comes from time constant C11/R23, following the rectifier based around U2 - which again receives its input signal from the input highpass/bandpass filters centered around U1 (that makes the sensitivity emphasis for the rectifier and attenuator)






check that signal here resembles what happens at ic U3 pins 1 and 14 are looking right, following hi-freq content

If it was me fixing this, I'd shotgun the three TL084's, starting with U3 - those jfet thingys don't like driving inductance. Make sure to socket, so you can experiment with alternative opamps.

/Jakob E.

PS: Posted the inspiration for this, the Ortofon STL-631 Cutting HF-limiter (de-esser) here: https://groupdiy.com/threads/ortofon.88483/
PPS: MagAmps are cool

 

[thelivingroom]

Upper L corner, 100uF:

"Added this to filter out AC
that was getting in this DC
line and wreaking havoc
"
..you don't want this extra electrolytic here, it messes seriously with your time constants and upsets the current-drive opamps. If there's excessive AC component here, it have to come from the sidechain or driver

Current through reactor/attenuator is differentially driven between 1) U3A, TL084 Pin1, adjusted through RV#3 and 2) driven from inverting amplifier U3D, pin14

this also drives panel meter

Input signal for all this comes from time constant C11/R23, following the rectifier based around U2 - which again receives its input signal from the input highpass/bandpass filters centered around U1 (that makes the sensitivity emphasis for the rectifier and attenuator)






check that signal here resembles what happens at ic U3 pins 1 and 14 are looking right, following hi-freq content

If it was me fixing this, I'd shotgun the three TL084's, starting with U3 - those jfet thingys don't like driving inductance. Make sure to socket, so you can experiment with alternative opamps.

/Jakob E.

PS: Posted the inspiration for this, the Ortofon STL-631 Cutting HF-limiter (de-esser) here: https://groupdiy.com/threads/ortofon.88483/
PPS: MagAmps are cool

Thank you so much! Right after I posted this I received an email from Manley that they found the hand written calibration instructions from 1995, so I’m about to pull that cap I added and calibrate it. The first thing I did was add sockets and change the three IC’s, but I’m going to plop in a new one in U3 again before calibrating in case I wrecked it. Fingers crossed that I’ve been chasing my tail this whole time for one misplaced resistor and an out of whack calibration from someone else’s prior work, but I’m happy it forced me to learn something new at least.

I initially thought this was a silly piece, but after reading up on reactors/mag amps it gave me a different perspective.

PS: your G21 is unique and amazing. I’d love to own one someday, along with everything else you make.

 

[thelivingroom]

Upper L corner, 100uF:

"Added this to filter out AC
that was getting in this DC
line and wreaking havoc
"
..you don't want this extra electrolytic here, it messes seriously with your time constants and upsets the current-drive opamps. If there's excessive AC component here, it have to come from the sidechain or driver

Current through reactor/attenuator is differentially driven between 1) U3A, TL084 Pin1, adjusted through RV#3 and 2) driven from inverting amplifier U3D, pin14

this also drives panel meter

Input signal for all this comes from time constant C11/R23, following the rectifier based around U2 - which again receives its input signal from the input highpass/bandpass filters centered around U1 (that makes the sensitivity emphasis for the rectifier and attenuator)






check that signal here resembles what happens at ic U3 pins 1 and 14 are looking right, following hi-freq content

If it was me fixing this, I'd shotgun the three TL084's, starting with U3 - those jfet thingys don't like driving inductance. Make sure to socket, so you can experiment with alternative opamps.

/Jakob E.

PS: Posted the inspiration for this, the Ortofon STL-631 Cutting HF-limiter (de-esser) here: https://groupdiy.com/threads/ortofon.88483/
PPS: MagAmps are cool

It may be the reactor, unfortunately. I still couldn’t get it to calibrate at all, so I set my DMM to measure current, inserted it between the input wire to the reactor and the junction of D4/D5 where it connects, and set my signal generator to do 20 second sweeps from 5-20kHz at 10dBu (the level specified in the cal notes) and the side chain is working, going from 15ish mA full CCW threshold down to 0mA full CW across numerous adjustments of the Hi-Mid and High pots. If I set the threshold at 12 o’clock and do the same, the current changes throughout the 5-20kHz sweep coincide with where I set them, as well. With a steady signal, adjusting RV3 and RV2 changes the current, so they’re functioning. With threshold full CCW, monitoring the output of the unit shows a .8ish vAC boost in amplitude over the 10dBu at the 5kHz start and gradually rolls off down to 430mVAC at 20kHz. My only thought is that maybe it’s supposed to get more than maximum 15mA of current that I can get out of it, but I’ve pulled a leg on every part and tested them, and nothing is busted. About to put the lid on it and put it back on the shelf haha.

 

[thelivingroom]

I dove into this, mapped out the side chain PCB, drew the schematic for it, and read about saturable reactors last weekend before starting to troubleshoot. It had two issues, the main one being that I noticed R33 (top right in the schematic) was for some strange reason connected to pin 13 and its associated trace, therefore bypassed by the trace and doing nothing. In the one somewhat clear picture I could find online, from a Reverb sale 7 months ago, I saw that it should be in the feedback loop between pins 13 and 14. I fixed that and the DC feed to the reactor now seemed correct when working the threshold (from 11ish volts full CCW down to 0 VDC at full CW), but it still had the crackle. I scoped every output pin on the TL084's and when I got to pin 14 that feeds the reactor, sure enough there was AC showing up in that line, probably from the PSU/line amp PCB that's directly behind the reactor. I increased cap values from that line to ground until it wasn't showing up on the scope anymore, and now it's clean and everything appears to work properly up to the reactor, but I'm still getting no reduction on output when testing it with a drum overhead track, vocals, or signal generator. I've tried different combinations of adjustments to RV2 & 3, numerous frequency ranges from 5k to 18k, sweeping all the settings, etc. My one fear is that maybe the reactor is toast, because I have found nothing online about it. Searched the TDK model number and a ton of combinations of other info that's printed on it and names that it may be called, and nothing. The resistance on the AC side of the reactor is measuring really low, like 2.2 Ohms, and I don't know if that's how it should be or not.

Maybe @gyraf has some insight?

I updated the schematic in this post, and also added the calibration notes to it.

 

[L´Andratté]

What are the reasons to use a "mag-amp" in this circuit?
Besides being a cool concept for electronics-nerds like us of course...

 

[gyraf]

Because there was not any other simple ways of doing a fast precise voltage-controlled filter

 

[thomasdf]

That’s a very intriguing design and unit, I wonder how it sounds ?!

Thanks for drawing the schematic in such a clear / readable / elegant way !

 

[thelivingroom]

That’s a very intriguing design and unit, I wonder how it sounds ?!

Thanks for drawing the schematic in such a clear / readable / elegant way !

I’m wondering how it sounds too, but I’m taking a break from it for a few days!

Thanks for the compliment; for the past 8 months or so I’ve really focused on trying to learn less glamorous than building gear helpful things that I should have long ago, like KiCAD, Illustrator for better front panel graphics, and mapping out PCB’s that I’ve never been able to find schematics for. I still have an infinite ways to go with learning/understanding electronics in general, but it’s all tying into helping me improve. I guess I like going backwards haha.