[_scotty_]

Nathan, have you been able to apply a signal  to R405 or R403 to achieve a direct path to the input of the Tripath module. Also have you figured out how to lower the power supply impedance at the amps switching frequency.

[Occam]

Bypassing the input card does not result in a phase mismatch at the output in my modifications.  Mono/bridged mode is achieved by sending the signal to BOTH CH1 and CH2 inputs, and ground to one of the available input grounds.  Both setups have been tested and prove functional and incredibly positive.  Both setups have also been tested with a phase tester and maintain absolute polarity from input to output.

I hope this helps a bit.

Nathan

Yes! Thank you!

The 'extra' inversion for the single channel is provided by reversing the output connections in one of its loudspeaker ouputs.

This is not apparent from the schematic. This only becomes apparent, for those of us without the actual amp, by looking at the loudspeaker output connectors on the back panel. You can see it by downloading the manual (upper right corner of the CarverPro page) -
http://www.carverpro.com/2003/products/zramps.html

and noting which posts are used for bridging (page 10 of the ZR1000/1600 manual) and that the binding posts are not (+) and (ground), but rather (+) and (-). Channel 2's (-) ouput is ground and Channel 1's (+) output is ground. While this is not  the 'normal' connection, it is very clever in that it implements Tripath's reccomendation of running one of its chips channel's inverted from a previous stage, and provide the requisite reversal with the loudspeaker connection. In typical stereo reproduction, with a large amount of in phase material, it makes efficient use of the mono, shared power supply, minimizing power rail droop.

[8thnerve]

Nathan, have you been able to apply a signal  to R405 or R403 to achieve a direct path to the input of the Tripath module. Also have you figured out how to lower the power supply impedance at the amps switching frequency.

Honestly, I haven't had time to do anything but sleep and mod.  I have a pile of these amps to do, with more arriving every day!  Once the initial frenzy dies down, I may have time to play with the unit more, but for now it's work, work, work.

[Occam]

We now know the following as to the ZR amplifier's architecture-

A gain block consisting of the circuitry on page 3 of the schematic provides -

1. Voltage gain
2. Pro features - high pass, low pass, level control, limiting, as well as the drive for the signal present indicator.
3. One of these channels, CH2, inverts the signal presented to it from the Input Board, and compensates by inverting that speaker's output wiring.
4. The terminal opamp for each channel also provides low  pass filtering to minimize above audio band signals.

Whether we use those pro features or not, each channels signal will pass through 5 opamps and 2 coupling (and their bypass) capacitors.

We also know that on the Input Board, as it provides bridgeing capability as well as dual mono on Channel2, that at least for channel 2, and possibly for Channel1 as well, it can output (with a push of the mono button) both phases of its input signal.

So what does one do in order to bypass these 5 signal path (& 2 ancilliary) opamps per channel and take the Input Boards signal directly to R246 & R247?
We do it ass backwards, of course! -
Which is to say that we derive our [EDIT(mono brigding signals for each channel from the same phase on CH2) This is incorect, bridging will still require both phases of CH2], signal for stereo operation, we take the reversed phase (in relation to CH1) to driveCH2 on the Tripath chip. We know that the requisite 'lost' phase signals are produced on the Input Board, and are most likely produced by an inverting voltage follower.

Which leaves the gain loss and 'above sonic' filtering action.....

the gain loss might be provided by the a change of the Input Board's, gain jumpers, feedback resistors, or an increase in volume setting on the source component, or some combination of all three. As all opamps, whether on the main or input board, are fed the same + & - 15volt rails, we know this can be done. Or we can change the gain of the Tripath chip itself, which would involve changing the gain setting, input to the Tripath chip, resistor. I believe that on the Tripath chips used in the ZR amps, the chips feedback resitors are internal to that chip, so its gain adjustment would only be through the input resistor, which is located under the heatsink assembly....

What filtering (that is found neccessary) might be actively provided via the Input Board's input differential amplifier, or by additional passive ciruitry. This needs further exploration.

And obviously, the opamps that reamain in the signal path might benefit from a replacement. There are scads of pin compatible, dual opamps available. The extent to which additional powersupply regulation or decoupling would be needed,  and their benefits, would depend on the opamps chosen.

[_scotty_]

I hate op-amps. Rant deleted.

[Occam]

I hate op-amps. Rant deleted.

Scotty,

I rather enjoyed your rant.... When I see an opamp, I spit.

So I'll address this post to the 1 opamp/channel that one can't escape on analog input Tripath chips, the one that is internal on the chip, that accepts the input signal.

This opamp is a single rail opamp that runs of a +5volt rail. It runs as an inverting amp with an externally provided approximately 2.5volt offset, and in Carverpro's implementation is provided by an opamp servos. ICs U403 & U404.

Although the tripath chips, TA0103a & TA104a, are 'end of life' products and are no longer listed on the Tripath site -
http://www.tripath.com/audio.htm

One can discern the TA0104a's configuration from the TA0105a PDF, page 23 & 14.
The voltage gain of this chip is from 2 components withing the Tripath chip, the inverting opamp and the modulator. The modulator gain on the 104a chip is fixed (I believe) at 40, and the input opamps gain is determined by the internal feedback and input resistiors, as well as any added input resistor. The fixed feedback resistor of 20kohms, and the input resistors, 5kohms plus the external input resistor, Ri, determine the input opamp's gain by the standard inverting formula -

Av = -20,000/(5,000 + Ri)

and the initial highpass pole is a standard 1st order filter determined by the input impedance, Ri + the 5,000ohms internal resistor, seen by the blocking capacitor,  C420/425, .47uf.

Fp(hz) = 1/( 2 x Pi x (5000 + Ri) x C)


Changing the external input resistor -

If one wishes to bypass that 'pro' feature circuit, in order to eliminate 5 opamps and 2 capacitors in the signal path / channel, one can simply change the external input resistors, R405 & R403. Realize that lowering the external input resistors will increase the gain of the Tripath's input stage, compensating for the gain lost in bypassing that 'Pro' circuitry, and will also change the input impedence of the Tripath as well as the pole frequency! The issues of adding the extra inversion of one channel (lost in the bypass surgurey) will be addressed when the schematic for the Input Board is made available....

Changing the resistors R403 & R405 from 100,000 (on the ZR1600) to 2,500ohms will increase the gain of the input stage from -0.1905 to -2.6667, almost exactly compensating for the gain loss of bypassing the 'pro' circuitry. It will also raise the pole frequency,  the blocking caps seeing the full 7,500ohm input resistance,  to 45hz. One would also have to change the input Caps, C420 & C425 to lower the pole frequency. A 4.7uf BlackGate NX is not overly expensive, and would give a pole of 4.5hz.
As the Input board is configured to drive a following input impedance of 2.2kohms, the 'new' impedance of 5kohms should not be problematic.

The above formulas should allow the modder to adjust the gain of the Tripath chip, as well as calculate the input impedance and pole frequency.
For those who wish to drive the Tripath chip on the ZR amps directly, one must carefully consider the output characteristics of the source component. For those with passive preamps, the typically low impedance of the (for noise purposes ) passive pre, in paralell with the lowered impedance of the modified Tripath Chip's input, may find it problematic.

When the schematic for the Input Board is made available, I'll try make this sensible....

[DAVID GARLETT]

We have had many request for line drawings of the  main PCB board. We have added them to the schematic file already availible on the site. If anyone has trouble downloading, as always your welcome to call and have it emailed direct....They should be availible this afternoon.

Thanks
David Garlett

[Cens]

Any of you gents playing with the ZR1600 have an opinion on the pros and cons of adding a tube (6922 or 6DJ8) buffer stage at the input and wiring directly to the tripath chip?

Regards,

Chris

[_scotty_]

The Tripath module does not need a buffer of any kind ahead of it for the amp to function properly.  The modifications discussed so far have been an exercise in trying to eliminate anything between the input setting gain resistors and the input jacks on the back of the amp. If the sound of tubes is desired in the system a tube preamp offers more options and doesn't
add additional complexity to the amplifier.

[Cens]

The modifications discussed so far have been an exercise in trying to eliminate anything between the input setting gain resistors and the input jacks on the back of the amp.

Thanks.  The thinking behind the tube buffer idea is not that it would be necessary for the amp to work -- I agree that it isn't -- but could allow for additional flexibility.  Per my admittedly inexpert understanding, by bypassing all the parts associated with the pro elements, you also make the amp somewhat more difficult to drive without an active preamp.  If I wanted to run the ZR1600 with a passive or direct from a CD player with volume control, for instance, I'm guessing that the buffer stage would help with impedence problems.  Does this sound reasonable?

Perhaps more important to me, I like the idea of a tube-tripath hybrid amp.  Although any sonic improvements would be entirely in the ear of the listener, it could be an interesting mix.  (Or, of course, it could suck.)

Regards,

Chris

[_scotty_]

cens, I use an active buffer as a preamp and the buffer drives the 20kohm
input impedance of my amp.  An impedance matching buffer of some sort internal to the amp would allow a source to directly drive the amp.  The volume pot would be between the the source and the buffer.  I have now
in a round about way described a single input integrated amp at this point.
A passive preamp is never a good a idea.  The concept was embraced as a result of the admittedly pathetic performance of any number of active preamps.  The passives' only advantage is that it sounds better than a poor sounding preamp.  With a properly designed and executed active circuit you will have better performance. I think the same sonic result
of a tube-Tripath amp would be acheived with a well executed zero gain tube buffer stage as a preamp, outboard of the ZR1600.  The ZR1600's performance would be uncompromised and you would still have the option
trying something else later.  With a tube circuit married to the ZR1600
changing your mind later would not involve surgical procedure on the ZR1600.

[Occam]

Any of you gents playing with the ZR1600 have an opinion on the pros and cons of adding a tube (6922 or 6DJ8) buffer stage at the input and wiring directly to the tripath chip?

Chris,
As Scotty has said, driving the Tripath chip directly with a 6dj8 buffer should not be problematic. But a number of parameters should be considered -

A tube buffer, I'm assuming a cathode follower, should be able to  drive, at minimum, a 20k load.  A 20k load would define a the gain of the Tripath's internal single rail opamp of 1, for an overall gain from the Tripath chip of 40. In order to drive the amplifier to full power (approx 4.5 volts peak to peak at the output of the internal opamp) your source component would have to be able to provide that full voltage swing as a cathode followers gain is one, at best. The standard 2volt RMS (2.8volts peak to peak) output of many CD/DVD players would not provide sufficient voltage swing to produce full output from the amp.

Increasing the input impedance of the Tripath chip to an easier load (for the tube buffer) to 50k, would impose even higher drive requirements. That 50kOhm input impedance would define the Tripath's opamp's gain as -0.4. This would impose a swing requirement, for full output, of 11.25volts peak to peak. This is of a magnitude not generally provided by opamp output source components. (refer to a previous post for the formulas for gain and pole frequency).

You also have the opportunity to use the output capacitor of the cathode follower as the required input capacitor on the tripath chip.... But unless your buffer provides a 'mute' until warmed up, you might find a rather large turn on thump. This can be minimized by the expediency of powering you buffer before the Trippath amp is turned on.

T'ain't no such thing as a free lunch...

Regards,
Occam

[Cens]

Scotty/Occam,

Thanks for your thoughts.  You've given me much to think over.

Regards,

Chris

[Occam]

And for those of you who don't know how to do dBV calculations.....

Bypassing the Input Board and that 'superfluous' pro feature circuitry [Taking 7 opamps/channel out of the signal path] and replacing the 100kOhm input resistors to the inputs of the Tripath, (and proper input cap values for an appropriate hipass pole) with a -

35kOhm input resistor, for a input impedance of 40kOhms, will yeild a a voltage gain of 20, which happens to be a gain of 26dB, which sorta seems to be a rather standard 'audiphile gain for a power amp.

or a

15kOhm resistor, yeilding a input impedance of 20Kohms, provides a gain of 40, which happens to be 32dB, which appears to be a high gain 'standard'. [or put a removeable 26.25kOhm in paralell with the above 35kOhm, and you've an adjustable gain amp]

So if your present preamp (of whatever topology) will adequately drive a 40k input impedance amp with a 'typical' 26db gain, be it digital or analog,
You're there....

and besides, if one's present preamp will drive the 2.21k load resulting from just bypassing the Input Board, it sure as heck is gonna drive the higher load of a directly driven "Tripath chip only" amp.

[PeteG]

I what to let everyone know of a problem I have with my upgraded ZR1600 amp. I upgraded the amp similar to Eighth Nerve but changed the wire from the main board to the output board too. So far it sounds very good (still breaking in) and the sibilance is gone so is the edgeness on the top end.

So to the problem, I've used the amp with three different speakers and it worked fine but when I hooked it up to my 4-Ohm subs it go's in protect mode every time and will not play, but hook it backup to the different speakers (8ohm-8ohm-6ohm) it plays fine.

I don't know if it's do to taking out the original fan (dual speed fan??).
Any idea guy's.


Edit:I put the carver back with my sub's and everything is fine, must have
been a bad connection I didn't see. One thing about a digital amp driving
the bottom end is it's very very clean and in control.

[JCC]

I just noticed that they have been added to the set of schematics on the Carver Professional web site.

[Monolith]

I just noticed that they have been added to the set of schematics on the Carver Professional web site.

Thanks for the update notice.  I've changed the link title on the first post of this thread to reflect that the single PDF file now contains the schematics, parts layout, and board traces.

[Occam]

A brief summary of calculations for the Tripath TA104a chip -

Voltage Gain = Gv = Rf/ (Ri + Rii) x Gm  = 20,000/(Ri + 5,000) x 40

Where Gm is the Tripath chip's modulator gain, which is fixed at 40 on this chip.
 Rf is the internal input opamp's feedback resitor of 20KOhms
 Rii is the internal (to the Tripath's input opamp) series input resistor, 5kOhms.
 Ri is the external series input series resistor, R403/405, which is the only user selectable component that allows adjustment of the Tripath chip's gain and input impedance.

The input impedance is simply Ri + 5,000

The voltage gain expressed in dBv is simply -

dBv = 20 x LOG10(Gv)

Tripath chip input highpass filters -

An input pole (dc blocker) is established by the series input resistors, Ri + 5,000, and the series input capacitor.
this pole frequency may be calculated as -
f = 159,155 / (R x C) where R is in Ohms, C is in uf, C420/425.
rearranging -
C = 159,155 / (R x f)
R = 159,155 / (C x f)

On to actual calculations using the above formulae.

On the stock ZR1600 (just looking at the Tripath 104a chip) -

Gv = Gain = Rf/(Rii + Ri) x Gm = 20,000/ (5,000 +100,000) x 40 =
.1905 x 40 = 7.62
in dBv = 20xLOG10(7.62) = 17.6

and the input impedance is I = Rii + Ri = 5,000 + 100,000 = 105,000 Ohms
and an input highpass frequency of f = 159,155/((Rii+Ri) x C) =
159,155/(105,000 x .47) = 3.23hz

Now lets decrease the input impedance from 105kOhms to a more reasonable 40kOhms, replacing the existing 100k R403/405 with 35kOhm resistors, which along with the internal 5k input resistor provide the 40kOhm imepdance and increase the gain to a more useable number. By the above formulas

Gv= 20,000/(35,000 + 5,000) x 40 = 20
dBv = 20 x LOG10(20) = 26dB

and the highpass pole with the existing .47uf coupling caps would now be

f = 159,155/(40,000*0.47) = 8.5hz
which is certainly higher that the original 3.2hz, but nevertheless, is a very reasonable input pole. I'd also try a 1.0uf cap here, for subjective evaluation.

Similarly, we could double the amps gain to 32dbV by using a 15kOhm external input resistor for a total input impedance of 20kOhms. This would also cut by half the required drive voltage for full output to approx 4.5v p-p. But the downside is the source component feeding a lower impedance and its current requirements....

[Occam]

If you look at Monolith's pictures of the +-15v regulators -
http://www.audiocircle.com/index.php?action=gallery;area=browse;album=122

and the +5v regulator -
http://www.audiocircle.com/index.php?action=gallery;area=browse;album=122

One will find JRC(NJM)7815A/7915A regulators, which are quite well specified, and are well thought of by many subjectivists.

Unfortuneately, the same cannot be said of the ST L7805. It is a noisey poorly specified part.  ST makes a far better regulator the L7805A, but I'd suggest a Motorola (On -Semi) TO-220 part -
http://www.onsemi.com/pub/Collateral/MC7800-D.PDF

(don't forget to use an insulator (mica, thermopad...) when mounting 3 pin regulators who's tab is 'hot'.)

Certainly, an adjustable regulator with the ability to capacitively bypass the voltage reference, or a discrete.... would possibly provide even greater improvements, but these are not drop in replacements.

As the 5volt regulator in the 'bypass all those crappy opamps mod' provides the voltages for that last remaining signal opamp, the one internal to the Tripath chip, as well as power to the digital processing, this is the place to splurge by spending $2 for an upgrade.

And yes, the power supply's decoupling caps. So I'll preemptively comment. Yes, the BubbleGum FkU type are the only thing to use. Unless, its the Johnson 6 pole electrolytics. I personally prefer good industrial low impedance electrolytics, but only if bypassed by film and foil caps wound nearby in Sunset Park, Brooklyn. These caps a wound on the inner thighs of some of the very few extant, 'Maidens' hereabouts, while the Reader holds forth with Victor Hugo. I've found those caps, wound to the cadence of Dickens to have foreshortened depth....

[Occam]

6Mooms has just relased a review of EVo's Tripath based amp in their "latest bestest' iteration. -
http://www.6moons.com/audioreviews/bcdgen2/gen2.html

The single change was the replacement of the industrial , metalized polyester cap/channel on the output filter, with an industrial polyprop cap, which yielded (per the review) substantial subjective improvements.

While the output filter on the Zr appears substantially more complex than the EVo's, this does give some indication that  the ZR might benefit from enhancement. The Carver output filter has multiple poles, and the caps used are bypassed.....  tweeking this section is not for the feint of heart, and one would hope the modder would minimally have access to an oscilloscope that can be used as a spectrum analyzer, and know how to use it.

From the color, description, and origin, my WAG is that these replacement caps are either Rifa PHE450(427) or Vishav MKP1841[M](1846). The parenthetical designations are the prior series numbers. These caps are specifically taylored for the task, shunting high current, high frequency noise. This is a situation where the audio community benefits from the economics of SMPS and CRT monitor production...


Based upon specifcations alone, which may have no bearing on subjective qualities I'd start by replacing the .1uf filter caps with PHE450 630VDC/400VAC model #PHE450MB6100JR06. One would have to evaluate the necessity of of keepiing, or upgrading, the associated, extant bypass caps. No, I've no idea as to where you could source these caps.