how do i get OPA134/lm6171 to replace a TL071 in trident?

[rjd2]

hi folks, im in the process of upgrading the channels on a trident fleximix, and tried swapping some IC's out, to no avail. when i drop-in an OPA134 or LM6171, i get a noticeable drop in gain, and the EQ starts oscillating when engaged. i've tried to read up on this both here and in the GS archives, but im obviously not doing something right. i put .1uf ceramic caps from pins 4(V-) and 7(V+) to the earth(pin 5 of main pcb), but this didnt work. can someone explicitly spell out exactly what i would need to do to each IC locally, by chance? do i need a resistor or something across the n/c pins? i understand that a .1uf ceramic cap is to be put from each power pin to "ground", but its a bit hard to assess the "ground" point. soldering im fine, conceptually, a bit in over my head here. thanks much-rj

ps-a link to the schems:

http://www.studiomanuals.com/docs/tride ... _schem.pdf

[Nick Sevilla]

Hi,

Did you read this?:

http://www.ti.com/product/tl071

"TLE2071 - TLE207x offer the highest precision upgrade"

TLE2071

"The TLE207x series of JFET-input operational amplifiers more than double the bandwidth and triple the slew rate of the TL07x and TL08x families of BiFET operational amplifiers. "

I would try using these as replacements instead.

Cheers

[The Scum]

Which ICs have you changed?

Which parts, on which pages of that schematic?

[rjd2]

page 4, which is the "channel" module. i changed IC's 1&2 to OPA134's, and IC's 3-7 to LM6171's. when i changed IC 1&2 back to tl071's, the EQ IC's, 3&4, still caused oscillations.

fyi, at the 100pcs price point, the price on the TLE2071's comes down to .73/piece, which isnt bad. i may just do those, as they are drop-in replacements.

thanks for the help, folks.

[Jim Williams]

TLE's are not drop in replacements. They are an improved design, but have much greater bandwidth and different phase margins.

In order to successfully install a 100 mhz bandwidth part in a sloppy 2 mhz bandwidth design, you are going to have some issues. In addition you also have several "design" flaws to contend with, like opamps that were converted to voltage followers but they left the low value gain shunt resistors in place causing an unnecessary load on those opamps = THD.

Gerald Graeme's "Optimizing Op Amp Performance" has most of the data you will need to become an ace at this, but you better be pretty good at math.

After all, it's called engineering.

[Nick Sevilla]

TLE's are not drop in replacements. They are an improved design, but have much greater bandwidth and different phase margins.

In order to successfully install a 100 mhz bandwidth part in a sloppy 2 mhz bandwidth design, you are going to have some issues. In addition you also have several "design" flaws to contend with, like opamps that were converted to voltage followers but they left the low value gain shunt resistors in place causing an unnecessary load on those opamps = THD.

Gerald Graeme's "Optimizing Op Amp Performance" has most of the data you will need to become an ace at this, but you better be pretty good at math.

After all, it's called engineering.

Thanks Jim, as usual, your genius at this is wonderful to read.

BTW I put back my experiment on my G&L 2800-32 console, where I had changed some 5532 opamps with something else. It was really not worth the effort.

Cheers

[Jim Williams]

If you had a Harrison or something high end, then you could plug those parts in and they would work.

Most of these low end designs are short loaded in parts that ensure stability. After all, if the chip works without those extra parts, why pay to put them in?

The last thing they ever envisioned was some smart a** 30 years later trying to stuff the proverbial square peg into the very round hole.

But being as we are, hopelessly inquisitive, some of us are bound to try something whether or not we fail or succeed.

This is human nature, it's how we got out of those damned caves.

Anyway, looking at the rear of a successfully modifed Trident Flexmix input strip, there are many small parts added to the rear of the circuit board, many resistor values are changed. The last set I sent out was loaded with National LME49990 opamps, those are super quiet, super low THD. They also sucked 10 ma each! I used very large psu filter caps, machined sockets and most all the electrolytic signal coupling caps were yanked except in the EQ. Those little EQ modules were rebuilt with Dale resistors and Wima/Rel-Cap poly film caps. I reset some of the frequencies. The mic pre card was redesigned, it's the worst IC/transformer design Bill Whitlock at Jensen says he's ever seen. In all, it's quite a bit of rework to get those decent sounding, all those mods took several years to accumulate.

It was also a lot of work to get out of those damned caves.

[rjd2]

ok, thanks for the help. unfortunately, i already ordered enough TLE2071's to rechip all the channels. if dropped-in, what would be the result? i understand, the board isnt going to magically turn into "pro level", that's fine with me. my intent is to find a nice happy medium between downtime on the console, and improvements in sound. should i just put the 071's back in, and keep it moving? thanks for any advice.

[Jim Williams]

More than likely it will oscillate from instabilities. You must use a scope to observe any chip swaps to determine if it's operating correctly.

If this was easy, I probably wouldn't have this job.

[dfuruta]

ok, thanks for the help. unfortunately, i already ordered enough TLE2071's to rechip all the channels. if dropped-in, what would be the result? i understand, the board isnt going to magically turn into "pro level", that's fine with me. my intent is to find a nice happy medium between downtime on the console, and improvements in sound. should i just put the 071's back in, and keep it moving? thanks for any advice.

In that schematic on page 4, what is going on with ICs 6 and 7?? Are the inputs labeled in reverse on the schematic?

A few things you might try re: stability:
In addition to the .1uF ceramic caps, try putting a 10uF polarized cap from each power pin to ground. Instead of putting the .1uF from rail -> ground, you could try it across V+ and V-.

Put a small (20pF?) capacitor across Out and In- on IC5 (and the other ICs if they're oscillating too).

Do you have an oscilloscope?

[Jim Williams]

There are more mistakes on those schematics than I can count. Consider it a travel brochure instead of a roadmap.

[dfuruta]

That one just jumped out at me. Pure positive feedback!

[rjd2]

thanks for the help, folks. FYI, the TLE2071's seem to work as drop-in replacements; no oscillations or issues im noticing so far.

[leigh]

TLE's are not drop in replacements. They are an improved design, but have much greater bandwidth and different phase margins.

In order to successfully install a 100 mhz bandwidth part in a sloppy 2 mhz bandwidth design, you are going to have some issues.

The datasheet I'm looking at for TLE2071's says they are around 10 MHz bandwidth, not 100 MHz. Am I missing something? Or perhaps you were just making a generalized statement?

My Trident (a Series 65, which also uses TL071's) has had many channels upgraded with either Linear Technology LT1357's or National LME49710's. Those are spec'd at 25 MHz and 55 MHz, respectively. I didn't make those upgrades myself, they were done before I got the board. But I don't think there were a ton of other mods done to those channels to ensure stability with the new parts. Should I haul out the scope and start looking for trouble spots?

I'm also planning on upgrading the rest of the channels soon myself, so I'd like to be mindful of potential trouble when picking the replacement chips for those.

thanks,
Leigh

[Jim Williams]

Got scope?

You'll need one.

The LM6171 as mentioned in the original post is a 100 mhz part. When stable it sounds good but is a bit noisy.