[Glassman]

enjoy

this is for my friend's Kenwood DP-7090 equiped with total of 8pcs of BB PCM1702 DACs and now with 12pcs of OPA627 on dual DIP8 to SIP8 reductions.. we'd like to find out how does Gilmore compare to various opamps, so that's the reason for this board

you have to stack two of these boards to make dual opamp to replace the original NJM4580L (btw. also found on Terratec cards and Audiotrak Prodigy 7.1).. so I'll have to etch 12pcs of this board to get six dual opamps.. this is gonna be fun

servo opamp circuit is left out as well as half of the 3rd stage (4 trannies instead of 8) and those 3k3 resistors from bases of the 3rd stage's trannies to ground.. just becouse there is no ground connection

Kevin say those prevent the bank from total shutdown of either PNP or NPN bank, but I suppose this is only in clipping situation which shouldn't happen inside the player.. there is always up to 2Vrms, powered with +/- 9V there should be no such clipping situation..

any comments?

 

[andrzejpw]

Wow, that's seriously cool! Let us know how it turns out.

 

[JohnFerrier]

Yep, nice work Glassman.

I guess you are using the 2SC2712 2SA1162 SMT devices. Maybe add bypass capacitors on the +/-9V lines. Also, it may have been better to put the more sensitive input circuitry closer to the connector. The input lines would be shorter and possible noise the outputs cause in the PS don't pass downstream to the input devices. I don't know if the smaller SMT run hotter with Kevin's bias points. Do you want to use a potentiometer to adjust the offset (matching SMTs will be a bit more challenging)? These are minor points though.

Great idea.


JF

 

[kevin gilmore]

Not to throw any snow on anyones sat dish, but my design
is a low open loop gain design and is not going to be
suitable for current to voltage converters. Running on
+/-9 you are going to have to change all the bias
resistors. You will still need a dc offset pot. You can
certainly try increasing the open loop gain by changing
the resistors in the first and second stages...

Its snowing like crazy in chicago

 

[Glassman]

Quote:

Originally posted by JohnFerrier Yep, nice work Glassman. I guess you are using the 2SC2712 2SA1162 SMT devices. Maybe add bypass capacitors on the +/-9V lines. Also, it may have been better to put the more sensitive input circuitry closer to the connector. The input lines would be shorter and possible noise the outputs cause in the PS don't pass downstream to the input devices. I don't know if the smaller SMT run hotter with Kevin's bias points. Do you want to use a potentiometer to adjust the offset (matching SMTs will be a bit more challenging)? These are minor points though. Great idea. JF

thanks for your comments..

I'll be using BC850 & BC860 as they are more common in Europe.. the problem with bypassing power lines is that there is no ground potencial.. maybe I'll just run a piece of wire somewhere on the kenwood's pcb to ground it, then I'll be able to add some NPO smt bypass caps.. placing input circuitry towards the connector is definitely a good idea, but as you can see, there is no chance running the output to the opposite side of the board.. as for the bias points, almost all resistors will need recalculation for lower supply voltage and also much lower output current requirements.. I don't wanna fry those SMTs

also I decided to left out even the potenciometer, I'll just try many LEDs and keep the one which will result in the lowest offset.. I hope it wouldn't drift a lot..

btw. I've just found one mistake

I will run separate power lines for input circuitry and the rest of the circuit, this should prevent output stage from interfering the input..

 

[Glassman]

Quote:

Originally posted by kevin gilmore Not to throw any snow on anyones sat dish, but my design is a low open loop gain design and is not going to be suitable for current to voltage converters. Running on +/-9 you are going to have to change all the bias resistors. You will still need a dc offset pot. You can certainly try increasing the open loop gain by changing the resistors in the first and second stages... Its snowing like crazy in chicago

I'm happy to see your reply, Kevin..

we will try this module not just in the I/V position but also at the output as a buffer.. this should be the best place - input is already voltage (from OPA627 I/V) and it will drive the line excelently with its low output impedance.. I admit, it would be better to add little offset pot..

can you, please, advise me how to calculate the right resistor values and how to set higher open loop gain? I'm definitely not an expert in this area, my friend knows what to do, I'm just designing the PCB

 

[ppl]

Kevin is Correct an op amp intended for I/v Converter apps require High open loop gain. I have noticed that allot of op amp rollers use Video grade op amps like the AD-825 in I/v converters. This type of application requires that the op amp have low noise and High open loop gain so as to obtain the Full resolution of 16 bit and greater DA converters. However i do like the idea of a discrete component design plug and play op amp.

Yes Kevin it’s snowing for the last three days in the northwest also. Now three days of snow hear generally is looked at as an emergency. Now rain No prob but more than 8 hours of Snow! OMG call the National Guard

 

[JohnFerrier]

No comments on converting to an I/V converter. Otherwise, all the bias points are based on the 1.6V of the LEDs. This should remain about 1.6 even with 9V. Maybe the LED voltage will be a little less, that will mean the transistors are not biased quite so much. The 7.5k adjusts this, of course.

Of course, it's in Kevin's description but:
(1.6V - 0.6V) / 500 ohms for 2mA; that splits into
1mA into 5k ohms for 5V; to furnish
(5V - 0.7V) / 1k ohms for 4.3mA; to
4.3mA * 510 ohms for 2.2Vdc; and finally
(2.20 - (0.6 + 0.6)) / (25 + 25 ohms) = 20mA (x 4).

An elegant design and nice choice of components.

Another concern comes to mind, the paper describes using a dual tracking power supply. External circuitry (on the same power supply) may interfere with your SIP amplifier. For example, if the supplies are not stable (a motor runs--I don't know if this is a CD player or something) then the output offset may change (esp. w/o the opamp for DC adjust). For best matching, you may want to try doing it in circuit (without the second stage installed). (I read on another forum where one guy is using +/- voltage regulators for each op-amp in his design...)

Also, I notice that the resistor (1k) that connects to the -I input (and is involved with setting the gain) also connects to ground.

Good luck.


JF

 

[Glassman]

Quote:

Originally posted by ppl Kevin is Correct an op amp intended for I/v Converter apps require High open loop gain. I have noticed that allot of op amp rollers use Video grade op amps like the AD-825 in I/v converters. This type of application requires that the op amp have low noise and High open loop gain so as to obtain the Full resolution of 16 bit and greater DA converters. However i do like the idea of a discrete component design plug and play op amp.

ok, so it seems like it won't be good for I/V, but as the buffer opamp which sums the I/V opamps outputs it's good, isn't it? there are eight PCM1702 DACs, each of them uses one half of the dual opamp for I/V conversion, those are already equiped with OPA627, after them there are another two dual opamps summing the outputs of I/V and serving as a buffers.. this should be the right application for Gilmore, right?
Quote:

Originally posted by JohnFerrier No comments on converting to an I/V converter. Otherwise, all the bias points are based on the 1.6V of the LEDs. This should remain about 1.6 even with 9V. Maybe the LED voltage will be a little less, that will mean the transistors are not biased quite so much. The 7.5k adjusts this, of course. Of course, it's in Kevin's description but: (1.6V - 0.6V) / 500 ohms for 2mA; that splits into 1mA into 5k ohms for 5V; to furnish (5V - 0.7V) / 1k ohms for 4.3mA; to 4.3mA * 510 ohms for 2.2Vdc; and finally (2.20 - (0.6 + 0.6)) / (25 + 25 ohms) = 20mA (x 4). An elegant design and nice choice of components. Another concern comes to mind, the paper describes using a dual tracking power supply. External circuitry (on the same power supply) may interfere with your SIP amplifier. For example, if the supplies are not stable (a motor runs--I don't know if this is a CD player or something) then the output offset may change (esp. w/o the opamp for DC adjust). For best matching, you may want to try doing it in circuit (without the second stage installed). (I read on another forum where one guy is using +/- voltage regulators for each op-amp in his design...) Also, I notice that the resistor (1k) that connects to the -I input (and is involved with setting the gain) also connects to ground. Good luck. JF

thanks a lot for the formulas

I think the power supply for analog part of Kenwood is clean and has nothing to do with digital and/or transport, it's just for the opamps, so I think there wouldn't be any problem..

the 1k resistor together with the 10k from out to inverting in sets the closed loop gain, but this is already set on kenwood's board together with some LPF components..

 

[ppl]

Hi john. I have no fault with the circuit. Infact this really intrigues me as while this is not suitable for the use Glassman had in mind it nonetheless is something of worth and I am thinking a discrete component Op amp that will plug into a 8 pin dip as a line stage op amp. If such a thing was available I would surly test it in the PPA. Heck what a combo a PPA Al La Gilmore, Way cool. Also Glassman you might consider rail to rail bypass, I read an elantec app note once that stated to never bypass from rail to rail however it seams to work for LC audio.

 

[kevin gilmore]

latest nutty idea

http://www.bryston.ca/chrismemo/christopher.html

The third picture. Interesting version of balanced
(i.e. identical slew rates) output without being balanced.
Gain of 2. Which means that you can use an opamp front
end on +/-15 driving the output stage which is +/-30.
In pure class A form this has the possibility of being
ultra sweet. Careful matching of components is necessary.

I'm starting to synthesize it now...

 

[Glassman]

this is a very cool approach

it's almost made for the use in PPA, don't you think ppl?

btw. Kevin, seems like all your new projects has to be run out of at least +/- 30V, don't you wanna calm down a bit?

seriously, do you think this type of output circuit can be powered out of just 24V or 32V used in PPA?

also, do anyone know where to find some discrete opamp designs suitable for I/V conversion?

 

[kevin gilmore]

quote
btw. Kevin, seems like all your new projects has to be run out of at least +/- 30V, don't you wanna calm down a bit? seriously, do you think this type of output circuit can be powered out of just 24V or 32V used in PPA?


UMMMMM NO. A ground splitter at several amperes not
so easy. Furthermore the idea is to get a voltage swing
greater than you can get with +/-15 or +30. So you could
certainly get by with +/-24.....

 

[ppl]

Glassman> I am quite sure that the Design could be set up for 24 volts. however My PPA Dimond Buffer board is going to allow higher than 24 volts to be used. the 24 volt restriction is to acomidate the AD-8610 that is max at 26 Volts. but then smaller Idss op amp fets will drop more voltage and allow any op amp you want at 30 volts.
My last Discrete Headphone amp ran of +/- 50 volt rails. The true meaning of headroom... Hee... Hee.... no pun intended.

Hi Kevin. no i did not intend to turn the PPA into a Loudspeaker amp. the present application works I am just considering a discreet op amp to go along with my Discreet buffer. However an op amp buffer combo is perfect as a head amp and adding a discrete output stage should improve Dynamics I have been playing with the Je990E. BTW thanks for Z Link That output stage looks similer to Pass Stasis output stage. and yes if i had room on the PPA i would certinly entertain the idea and Cascode output stage. The cascode output stage as in my old cas1, Leach doubel barrel, and my own cascoded designs still ring sweet high frequency detail to my ears after long time of not listening to these and i want to get back to that. maybe i should just start with a cleen slate and do an all discrete PPA no op amp adaptors and all that **** also will discourage undersireable part changes.

 

[kevin gilmore]

No the bryston stage looks nothing like the stasis stage.
There is both a pnp and a npn output on both the top and
bottom drivers. Look again at picture #3...