Dyno my piano

[tomtom]

Do you have any idea (or more info.) about the famous Dyno preamp for Rhodes piano, everybody raves about it, I've done some research but not any technical data, schem or others....

 I use a fairly linear boost but the bass are not round enough and the high lack some crisp...
 Recording direct and you have too much distortion because of the peaks...


  I would love to have a great fat and crisp sound to record with.

[Paul Perry (Frostwave)]

http://www.fenderrhodes.com/history/dyno.html

no schems, but at least some tips. Looks like the dymo effect depends partly on having the piano set up & adjusted correctly (just like a guitar!)

[tomtom]

Yes, I've read about that too and I think it's a big part of the sound but those preamp seems quiet compare to the onboard trem preamp found stock plus it gives big bottom and some crystal clear highs I can't have stock...
 I know speakeasy makes clones of them, I may buy one...it doesn't seems complicated and it's powered by a 9v wallwart so I ask  :wink:

 Many people used to record direct with this preamp (chick corea and others)

[ErikMiller]

I have a Rhodes, and Jack Orman's MOSFET Booster really makes it sing:

http://www.muzique.com/schem/mosfet.htm

I'd try one of those, either self-built (Jack sells PCB's for them now) or in the form of the Catalinbread Chile Picoso.

[tomtom]

Yes, I use a mosfet boost plug into the passive output of the Rhodes but I can't have both clarity and fat bottom (the bass seems too metallic)

Thanks anyway !
 Tom

[zjokka]

This is an old thread, but dried out without much info.

If you look at the amount of commercial guitar fx of which the schematics ARE available, it seems strange that a Dyno-My-Piano stompbox hasn't been tried yet. Of course the Chuck Monte's company has gone out of business, so it's not a matter of copyright etc, I figure.

There are some Rhodes techs that have developped their own boxes. More of that can be found on fenderrhodes.com forum. But no clue there as to the contents of the Dyno-My-Piano. Currently an "exact copy" of the original Dyno is being sold on Ebay. They provide a little peak at the circuit. Leaves much to guessing, but does it really need 3 9V batteries? Does it really have to cost $329?   

However it cannot be that hard to make it yourself.
Will continue my search till I can make one, but this is long term. If I really have to buy an old Dyno to trace the circuit, maybe I will try to score one off ebay. Here are the pics. I put them in my image account to avoid losing them after the auction closes







ZJ

[Eb7+9]

I'd like to see that circuit - heard they work good on Wurly's ...

[R.G.]

If I really have to buy an old Dyno to trace the circuit, maybe I will try to score one off ebay.

That's determination! Good for you. Over time, I bought a number of pedals for reversing. If you buy from ebay and then resell it, you don't get hurt too badly. Sometimes you even make money, although that's not something to count on. I bought, for instance, a Big Cheese for reversing, along with others. Broken pedals are even better.

Here are the pics.

It's very close to being reversable from those pictures. Fortunately, it's a very loose PCB that can ...almost... be traced through the top side. I make it four transistors and a voltage regulator IC. The input and output leads are fairly obvious. Does it really need three 9V batteries?? Can't tell. Maybe, depends on how big the signals are in Rhodes.

Does it really have to cost $329?

Sure it does!! That's what the market will bear... 

[zjokka]

It's very close to being reversable from those pictures. Fortunately, it's a very loose PCB that can ...almost... be traced through the top side. I make it four transistors and a voltage regulator IC. The input and output leads are fairly obvious. Does it really need three 9V batteries?? Can't tell. Maybe, depends on how big the signals are in Rhodes.

I think the signal in the Rhodes is high impedance, comparable to a guitar. I read some technical data recently but now re-Googled in vain to find the harp impedance. It's got 73 magnetic pickups set up in a series of parallell groups. Each pickup about 6k



The Dyno preamp was added to the passive circuits of the Stage model, to upgrade them to the possiblilities of the powered and preamped Suitcase config. So maybe, I could start from that Suitcase schematic, which is available, and build from there?



All the way yonder to the right is indictatd 25V. Of course there's the vibrato and the preamp. But I do need some help reading this.

Does it really have to cost $329?
Sure it does!! That's what the market will bear... 

Sure sure, but they could also offer it also as a $80 diy kit. That's a matter of policy... these are the guys teching Donald Fagen's Rhodes and so, see vintage vibe (dot) com.

ZJ

--
btw, rg, thanks for all the work done.

[Bernardduur]

I've build Albert Kreuzers bass preamp for a friend who has a Rhodes and it sounds AWESOME!!!!!! The gain can give grit to the beautiful rhodes sound and the wide EQ makes it a great preamp.

[zjokka]

thanks on the bass preamp hint. Looked at Albert Kreuzer's website, and this is truely impressive. Will maybe try it in a stripped down version if nothing else goes. Still I did get the partial schematic for the dyno preamp, that is, without values. My experience is still too limited to know what is needed to fill them in....

There is both a block diagram and a schematic without values. Don't know to which extent the values in the above schematic (for the Suitcase Rhodes - the one with built in amp and speakers), the controls are basically the same.





This material should provide some kind of a kick start, no?

zj

[Eb7+9]

getting closer ... you need one more page from the patent ... 

[zjokka]

getting closer ... you need one more page from the patent ... 

you think? 
Can you really user Spice to fill in the values? Have you built yours yet?

ZJ

[Eb7+9]

you bet - that's what a typical "reversing" job would be like - my build's right around the corner now ...

...

here's a clue, the author addresses anyone wishing to reverse this work by saying :

"To those skilled in the art to which this invention relates, many changes in construction and widely differing embodyments and applications of the invention will suggest themselves without departing from the spirit and scope of the inventntion.  The disclosures and the descriptions herein are purely illustrative and are not intended to be in any sense limiting. "

once you realize what he's pointing to, you'll post the missing piece of data and we'll get on with it ...

~JC

[zjokka]

hey, man why so cryptic. there is something going on with this, but I don't know what. nobody even dares to post anything about this in any Rhodes forum,

you mean this?



let's get on with it indeed...

but er... how, downloaded spice and all -- but this is my first encounter with the program.

thanks stil,
ZJ

[Eb7+9]

great ...

preamp circuits have have by nature high bandwidth response exceeding the audio range (up to ft of transistors or Miller depending) unless some high-frequency shunting is inserted ... this shunting is necessary to reduce high f noise among other things - you don't see this in guitar tube amps because the OPT at the end of the chain does this for us but in stand-alone pre's you do need it ... this is what the author refers to when talking about ultra-sonic caps - these are C3, C9 and C14 in the circuit ... they short out Drain-Source at high frequency (passed 16kHz) - ignore those for now we'll put them back in at the end of analysis ...

next, redraw the whole circuit without those Ultra caps and the Limiting diodes to have a simplified view - showing the Overtone and Normal signal paths running side by side going to the mixer node ... first thing to notice is the phase of the signal as it goes through each channel - notice they both arrive out-of-phase at the mixer node relative to the input ... this is important for it says the signals add there as opposed to cancelling ... that means we can use the superposition principle and treat each channel individually and then sum the individual components at the output - not a biggie, just simplifies things ...

let's assume that Source bypass caps offer full bandwidth boosting on all stages where they're used and that they aren't high-shelf (treble boost) or anything like that - an assumption to go back to if the end response doesn't jive with the predictions produced by the subsequent steps ... doing this leaves us with tone circuits that are buffered/amplified by more-or-less "flat" gain recovery stages ... one way to look at tone circuits is as frequency selective impedances - they load the stage previous to it in an impedance sense - so let's assume for now that we're using gain stage with fairly high drive (ie low-Z drive) ...

ok, what does that leave us with ?? we got "flat" gain stages loaded by frequency selective networks ... let's look at them ...

the first tone circuit made up of R4, R5, R6, C6, C6 acts as a flat "notch" network if R4 and R6 are made equal ... the second tone circuit made up of R17, R18, R19, C10, C11 is our friendly RCA tone stack minus the Mid resistor - this gives us our typical guitar amp (Fender Tweed) tone response curve with the mid-scoop at it's minimum ... by now you can probably guess what's going on here : that if you line up two consecutive filter stages like these in a row and MATCH up the notch frequency of the first to the mid-scoop of the second that you will produce a frequency response curve as depicted minus what happens in the low-F and high-F ... we end up with the Overtone response which is a combined tone stack response - one yielding a much deeper notch than what the RCA circuit can offer ... the Normal channel signal mixes back the signal that is missing so there's no hole in the overall response ...

here's where the data comes in handy - the plot shows a -30dbv notch at 450Hz (thank you Mr. Barry !!)
note: dbv = 20 LOG10 (Vout/Vin)

"fudging" with Spice gave me a theoretical -33dbv @450Hz using the following values:

R4=R6=270k
R5=1k
C6=100pF
C5=0.005uF

replacing R5 by a 1.5Hy inductor (a common trick in filters) brought the notch down to -40dbv ... in case that's of use ... 

as for the tone stack I'm currently playing with the following values :

[roughly scaled from the 5f-A Fender Tweed values by a decade both ways - C's up, R's dn]

C10 (treb) 0.002u
C11 0.2uF
C12 0.2uF

R19 (slope) 10k
R18 (treb) 25k
R17 (bass) 100k

combining both stages together in series with the buffer/gain stages gives a combined profile (using guestimated ultra-sonic caps) that's looking pretty much like the graph ... so the assumption of flatish stages seems to hold - the AC response of the Overtone channel is done ...

Normal channel response is quite straightforward ... the plot shows a flat bandpass response with a lower cut-off at 100hz and a high cut-off, again, similar to the Overtone channel and dropping off somewhere between 16k and 20k ... in the schematic that frequency will be determined by a series of components - C7, R10 and to a degree C1 and C4

as you can see, the quote that I put forth above shows what the author had in mind when he wrote that - namely that any style of gain block can be inserted between the tone stages to buffer them as long as the phase and gain levels are respected in the end ... the second is a moot point because these are typically adjusted against noise and output needs ... so, in other words, instead of using jFET gain stages to do all this we could use op-amps or whatever - UNLESS the jFET stages impart clipping characteristics that are pleasing ... you can see it makes sense to assume the jFET gain stages are flat in response since then all the spectral shaping is done by the tone networks so we don't have to recalculate taking gain stage frequency response into account ...  the "clipping" aspect refers to the Dynamic properties of the circuit, these are treated separately from the Spectral aspects ...

basically that's about it ... just a matter of implementing the whole thing (gain stages and mixing) your way ... what I'm playing with at the moment are gain jFET stages that have Rs=1k and Rd=3k, it's safer to work with low Drain loads - otherwise the frequency response of the filter network will interact with the (higher) driving point impedance ... keeping Drain loads low helps keep the simulation analysis from shofting although that can be remedies in Spice again (I actually simulated with 3k drive to make sure) ...

after all this is done, you'll want to tweak in ultra-sonics that fit the drive impedance of your gain stages - if you're using jFET stages with light Drain loads you may need caps that approach 0.001uF or even 0.01uF - best thing is to try those out by ear starting with the front and moving towards the back ... similar to what you'd do with a jFET od circuit to curb the noise ...

otherwise you could use op-amp gain block but have to fit frequency limitations to obtain the overall freq response curves - not more work really - just a slightly different approach has to be taken ... since I like the MOJO of class-A circuitry I'm probably going with the jFET's myself ... 2n5484's or any low VgsOFF device will do - like 2n5457, J201 and other favorites ...

some extra notes:

- choose R1 = 1k and R8 = 10k to simplify loading picture there ... or R1 and R8 in that typical ratio ... then choose C4 appropriately (for at least 100hZ feedthrough)

- C2 is chosen by taste - as in Lemme's PU analysis page this cap brings down in frequency the natural PU response peak

- R12 is chosen once all the gain levels are sussed - it's a little circular at this point, but we are fudging after all ...


unless the biasing of the stages are intentionally non-optimized for headroom (to induce early clipping or something) then that should be about it - for as the author says, many implementations are possible in duplicating his Dyno response ... nice thing about all this is the potential of customizing as some Dyno's seem a little bright

ps. notice a n-channel jFET is drawn in the gate circuit, it should be p-channel ...

[zjokka]

thanks so much for this analysis. took some time to take this in...

preamp circuits have have by nature high bandwidth response exceeding the audio range (up to ft of transistors or Miller depending) unless some high-frequency shunting is inserted ... this shunting is necessary to reduce high f noise among other things - you don't see this in guitar tube amps because the OPT at the end of the chain does this for us but in stand-alone pre's you do need it ... this is what the author refers to when talking about ultra-sonic caps - these are C3, C9 and C14 in the circuit ... they short out Drain-Source at high frequency (passed 16kHz) - ignore those for now we'll put them back in at the end of analysis ...

I think I understand this. But this got me going back to the  R O G  F l i p s t e r I originally wanted to built (and still want to try) to check if as stand alone EQ unit it should have those ultrasonics. I think I found them in the 22nF caps. Is that correct?

next, redraw the whole circuit without those Ultra caps and the Limiting diodes to have a simplified view - showing the Overtone and Normal signal paths running side by side going to the mixer node ... first thing to notice is the phase of the signal as it goes through each channel - notice they both arrive out-of-phase at the mixer node relative to the input ... this is important for it says the signals add there as opposed to cancelling ... that means we can use the superposition principle and treat each channel individually and then sum the individual components at the output - not a biggie, just simplifies things ...

I've tried to simplify the drawing, removing diodes and ultrasonics as you said. Added also the values you suggested and removed the dotted lines indicating the blocks -- just to get a clearer view:looks even more like the Flipster !

ok, what does that leave us with ?? we got "flat" gain stages loaded by frequency selective networks ... let's look at them ...

the first tone circuit made up of R4, R5, R6, C6, C6 acts as a flat "notch" network if R4 and R6 are made equal ... the second tone circuit made up of R17, R18, R19, C10, C11 is our friendly RCA tone stack minus the Mid resistor - this gives us our typical guitar amp (Fender Tweed) tone response curve with the mid-scoop at it's minimum ... by now you can probably guess what's going on here : that if you line up two consecutive filter stages like these in a row and MATCH up the notch frequency of the first to the mid-scoop of the second that you will produce a frequency response curve as depicted minus what happens in the low-F and high-F ... we end up with the Overtone response which is a combined tone stack response - one yielding a much deeper notch than what the RCA circuit can offer ... the Normal channel signal mixes back the signal that is missing so there's no hole in the overall response ...

ok. I understand what happens but I don't see this in the schematic. I see the different tone stack but don't see how they come togeter in the mixer. Is that in fact the output network?

basically that's about it ... just a matter of implementing the whole thing (gain stages and mixing) your way ...

What is the lower part of the block schematic that is (contrary to the rest) spelled out in components. How does this relate to the mixer? Or is this part of the overtone network? Also don't understand why some of the gain stages are drawn as opamps and some as transistors.

The text gives the following values and functions referring to the block diagram:

20 =
21 = 741 opamp
22 = blocking cap
23 = feedback resistor
25 = excessive reverse bias protection
26 = 2N4126 PNP hfe of about 100
27 = 4.7uF
28 = 20k pot
29 = 47k

How do I incorporate these?

ps. notice a n-channel jFET is drawn in the gate circuit, it should be p-channel ...

Do you mean it requires both types? Don't really understand this properly, I'm afraid.

ZJ

[Eb7+9]

here's another redraw to explain what I'm talking about ...

the normal channel's on the bottom, the overtone on top, the front-end pre stage is common to both ... the mixing network is drawn separately also at the far right ... the UL caps short out the Drain and Source of each device - they're aren't drawn here ... btw, these are not signal caps as in the ROG you mention ...



about the jFET biasing resistor values - I'm using 1k and 3k in Source and Drain circuits throughout, that's just a ball-park value pair I chose out of convenience ... mainly (i) to guarantee good headroom specs and also (ii) to make it easier to estimate filter network values ... if you increase the Drain load to 4k7 or higher (keeping Source load at 1k) you'll get more gain (output) from your gain stages but at a proportional loss of headroom (and noise increase) - which could be KEY to the Dyno filter's Dynamic Response like I mentioned ... by staying with this ratio (setting the voltage gain of the stage to a fair degree) you could use 10k, 33k or 100k, 330k to get similar gain and headroom but then you need to do more math to get the same filter profiles ... just trying to initiate the overall curve-approximating process by using easier to work with inter-stage values ... one thing though - a 3k Drain load in the last stages may make setting R12 more difficult or impossible (we'll see) ... if that's the case it just means altering biasing in last gain stages of each channel to fit the mixer impedance level better ... not a biggie to fix ...

another thing in this regard ... there seems to be quite a bit of gain developed in this circuit and I don't see why the filter would need to provide much make-up gain above that needed to overcome the attenuation factors in both tone networks - especially in cases where the PU is directly fed to the input of a guitar or bass tube amp ... in my rig I bypass the preamp and go straight from my Rhodes' PU to an Ampeg V4, many people like Fender Twins - same idea ... naturally I'd like the keep the output of the Dyno set not too high above Instrument Level range if I can - a little boost is fine ... I'm inclined to ditch the volume control at the output (just stick a resistor) and operate all gain stages at more optimized gain (and noise) levels, setting the output through the filter controls alone ... maybe that's why the Dyno needs a noise gate ?!

the gain stage bypass caps are chosen at 22uF throughout also - these are standard "RCA/Fender" values ... along with the 1k Source resistor they give a roll-off at about 7hz ... I think that's well below any tone-control activity happening elsewhere in the circuit ...

What is the lower part of the block schematic that is (contrary to the rest) spelled out in components. How does this relate to the mixer? Or is this part of the overtone network? Also don't understand why some of the gain stages are drawn as opamps and some as transistors

the author is using old telephony symbols here - a triangle meant an amplifier - this was later applied to op-amps by sticking two input in the front ... but you're right, he's using the same symbol to mean different things in different places ... for example, the 741 op-amp is drawn wrong in the gate circuit but I understand what he means // he's sorta combining the old telephony symbol thinking with feedback op-amp thinking - it's right and wrong at the same time ... it's a patent (wtf!) ... also, the degenerative feedback is really referring to Source resistance - in Systems theory this is referred to as Localized negative feedback to distinguish with Global NFB - he didn't have to draw that into the block diagram but he did ... maybe it's to mislead readers, maybe not ...

maybe we'll talk about incorporating the Noise Gate part later - there's a few issues that need looking over in that circuit also ... for example, can you guess where it goes in the signal path ?? (another mislead in the block diagram perhaps ?!) ...

I wanted to see how the filter would sound in a dynamically clean implementation so I drew an alternate op-amp version of this circuit (a so-called op-amp simulation) as the author suggests you can ... this approach does away with all the biasing issues of the single-ended version but it won't have any of that good ol' class-A mojo ... the layout for this op-amp version takes about 25 strips on IC-stripboard ...  certainly I will build both - nature of the beast ...

I think the basic spirit and scope of the invention is embodied here ... with the notch filter landing smack right on middle-C, with the bass lobe exclussively boosts the side of the keyboard below middle-C and the other lobe exclussively boosts the side of the keyboard above middle-C it's almost like splittting the keyboard and setting the mix between the two sides through the treb/bass controls of the overtone filter ... I imagine anything that accomplishes that basic function should embody the dominant Dyno effect - following that it's just a matter of shifting (parametrizing) the notch location and lobe spreads to taste to get a customized response ...

ultra-sonic cap estimates coming shortly ...

[zjokka]

there seems to be quite a bit of gain developed in this circuit and I don't see why the filter would need to provide much make-up gain above that needed to overcome the attenuation factors in both tone networks - especially in cases where the PU is directly fed to the input of a guitar or bass tube amp ... in my rig I bypass the preamp and go straight from my Rhodes' PU to an Ampeg V4, many people like Fender Twins - same idea ... naturally I'd like the keep the output of the Dyno set not too high above Instrument Level range if I can - a little boost is fine ... I'm inclined to ditch the volume control at the output (just stick a resistor) and operate all gain stages at more optimized gain (and noise) levels, setting the output through the filter controls alone ... maybe that's why the Dyno needs a noise gate ?!

i'm very curious about your disstatisfaction with the suitcase preamp. There is of course this big contradiction: the standard config Stage + Twin with definite tubesound versus the ('full-version') Suitcase (full but solid state). It's quite a different sound but if your (also) a guitar man you're bound to prefer the tubes. I have a 1977 stage 73 myself, but in my current band I'm mostly playing guitar, so the the keyboad player uses the Rhodes on his solid state amp, it sounds terrible. When I plug it in to my BF-ed Super Reverb, it's magic. Do have to adjust settings radically, but still very well-defined, clear sound, with barking dynamics. With your amp, I'm not suprised you want  to straightjack it...

Why does it need a noise gate? From the time I spent reading Rhodes fora last year, I seem to remember that they are very susceptible to noise because the pickup field is so broad. If you look at the pictures from the VV board, notice how they completed wrap it in copper shielding. I'm just freewheeling here..

Another question still preventing me from having breadboarded the current design is the 15V. Is that really necessary? For the same reasons of noise, I was imaging a stompbox that I could jack between Rhodes and amp and would sit on the Rhodes - maybe not even a bypass on/off...

maybe we'll talk about incorporating the Noise Gate part later - there's a few issues that need looking over in that circuit also ... for example, can you guess where it goes in the signal path ?? (another mislead in the block diagram perhaps ?!) ...

Logically, I would say, through away the noise before you start amplifying anything. I saw it was located at the power source, but you removed it in your drawing. They would go where now is the 1k and ** cap, after the input cap -- is it superfluous because 1k and cap ** perfrom that function?

I think the basic spirit and scope of the invention is embodied here ... with the notch filter landing smack right on middle-C, with the bass lobe exclussively boosts the side of the keyboard below middle-C and the other lobe exclussively boosts the side of the keyboard above middle-C it's almost like splittting the keyboard and setting the mix between the two sides through the treb/bass controls of the overtone filter ... I imagine anything that accomplishes that basic function should embody the dominant Dyno effect - following that it's just a matter of shifting (parametrizing) the notch location and lobe spreads to taste to get a customized response ...

So if you really needed a dyno effect, you could wire one up with a ABY box, a mixer and and broadband eq?

Another question I was wondering: what make that a cap should be polarized in the design? Are the ultrasonics polystrene caps?

thanks so much for the redraw
ZJ

[zeerust2000]

I've just stumbled on this site and read this thread with interest since I have recently aquired a Rhodes with a working dyno preamp which seems to have been wired up recently by someone with access to the schematic. At present I'm tracing the schematic to build myself a clone to fiddle around with.  Would anyone here be interested in more details of my unit? I can provide component values if you like.  The schematic matches the one in the patent except that Q2 is oriented differently and has a cap across it like the others.