Anyone have any recommendations for upgrades on an Adcom GTP-400. Yes, I know there are better pre-amps out there, but I have this one, and I'm not cash heavy to get something sweet. It sounds good and serves my purpose. Like most people on this forum, I'm capable of whatever tear down is necessary, so I'm open to suggestions. Any helpful input is welcome.
Thanks guys.
Thanks guys.
That site's blocked in China so I can't get it. I found an image with Yahoo but including it on this thread might violate copyright. Hence it'll be best for the OP to take his own - this has the advantage of making the thread self-contained. Threads with external links don't age so gracefully 
Right off...I would suggest locating the audio output op-amps and changing them to something newer or installing sockets so you can try what you want.
They appear to be IC553 and IC556, but it's hard for me to tell since the schematic isn't real clear.
Try a couple of OPA134's or any other DIP-8 single op-amp that floats your boat.
They appear to be IC553 and IC556, but it's hard for me to tell since the schematic isn't real clear.
Try a couple of OPA134's or any other DIP-8 single op-amp that floats your boat.
Looks like they used Adcom-branded parts where the usual model names would have seemed too profane - even for the LM317/337 regulators (low-noise circuit btw), now that's what I call some effort.
The designer must really have known their stuff - they even used Class A biasing on the opamps.
Noise, by contrast, appears to have been less of a concern than distortion, as the voltage amp (IC551/552) uses a 22k/3k9 feedback network. Apparently ADCOM 6A was an LT1056? Those would be JFET input parts, somewhat noisy buggers (14 nV/sqrt(Hz)). Sort of an LF356 upgrade option, methinks.
If the layout doesn't look too sloppy, I would give a good ol' NE5534 a shot, with the feedback resistors down to 5k6/1k. That should bring noise down a good deal. OPA132 should also work pretty well though, if somewhat more noisily.
I agree the schematic is messy. Pretty but messy. Oh, and it looks like the tuner will never be the last word in FM selectivity because the IF strip just isn't linear enough (loads of limiter amps, and limiting = nonlinearity by definition). Swapping the first IF filter X151 for, say, a 180 kHz job could help a great deal though.
The designer must really have known their stuff - they even used Class A biasing on the opamps.
Noise, by contrast, appears to have been less of a concern than distortion, as the voltage amp (IC551/552) uses a 22k/3k9 feedback network. Apparently ADCOM 6A was an LT1056? Those would be JFET input parts, somewhat noisy buggers (14 nV/sqrt(Hz)). Sort of an LF356 upgrade option, methinks.
If the layout doesn't look too sloppy, I would give a good ol' NE5534 a shot, with the feedback resistors down to 5k6/1k. That should bring noise down a good deal. OPA132 should also work pretty well though, if somewhat more noisily.
I agree the schematic is messy. Pretty but messy. Oh, and it looks like the tuner will never be the last word in FM selectivity because the IF strip just isn't linear enough (loads of limiter amps, and limiting = nonlinearity by definition). Swapping the first IF filter X151 for, say, a 180 kHz job could help a great deal though.
If you reduce the feedback resistance sure you'll reduce the thermal noise, but you'll also increase the dynamic noise as driving that impedance will give more switching noise on the supply rails. I've found dynamic noise reduces the excitement factor when listening, so I'd suggest going the other way - increasing the feedback impedances and if necessary substituting JFET opamps for bipolars so the opamp's own noise contribution (current noise) isn't an issue. OPA132 would be a good choice as already suggested.
14nV/rtHz isn't really an issue if your source is CD and levels are standard CD levels (2VRMS) as this in theory gives an audio band SNR around 120dB - some 27dB better than the RBCD format itself.
14nV/rtHz isn't really an issue if your source is CD and levels are standard CD levels (2VRMS) as this in theory gives an audio band SNR around 120dB - some 27dB better than the RBCD format itself.
Uh, but that's the gain stage after the volume pot, which tends to be kinda important in terms of noise. Its output noise sims at ~15 µV (not even taking into account whatever the following tone control may add). With a power amp having a 1.5 V input sensitivity, we're talking 100 dB, tops, and more importantly about 300 µV of output noise for a power amp gain of 26 dB. That's decidedly average, and while tolerable, would be quite easily audible on speakers of, say, 88 dB / 2.83 V sensitivity.
Given the generous SE Class A bias (7k5 to -18 V), I see no reason to be concerned about rail noise. This would, in theory, do for driving loads of a kOhm or less at 1.5 Vrms. Surely 6k6 shouldn't be much of an issue then?
I take it you're referring to the schematic which I don't have access to.
Yes the problem here seems to be not using the full dynamic range of the opamps by virtue of feeding their output into a 1.5V sensitive poweramp. For this reason I'll always use a step-down transformer when interfacing my opamp circuitry to my poweramp.
WIth resistive bias the rail noise is 'cleaner' than for a classAB stage (by cleaner I mean a less distorted version of the signal). However the rail noise still exists as the current draw isn't signal invariant - only with balanced CCS-loaded outputs does it drop to virtually zero. It would be best to run a sim to see what levels it turns up at, given the regulators and decoupling used. Sometimes the noise on the rail itself isn't degrading the SQ but the decoupling arrangement couples that noise into the 0V and that coupling to 0V causes the SQ degradation.
According to an article in Stereophile, Walt Jung and Adcom's product-development director C. Victor Campos designed the GTP-400.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.