Keyclick filters in T-series

[T-200]

Hi again T-club,</P>


This is just "thinking out loud" towrite downsome thoughts that I have been trying to figure out in my not so electrically genius mind...</P>


Are there any more click filters in T-series than the ones on the busamps and the rec/non vib amp (which I have fixed).</P>


I have put the small cap under the vibrato board as Carsten's advice was and now I have more leakage and keyclick when the vibrato is on, but when I turn the vibrato off the sound is less clicky and a bit dull. Does the vib amp somehow bypass the recovery amp? I also have a separate switch for lower vibrato on and that, too affects the upper manual sound. I haven't found a schematic with that switch and there are some newly added resistors connected between that and the "upper vibrato on" switch. Has anyone got schematics or pictures of the tab panel?</P>


I got an old copy of the schematics with my T-200 percussion board and somebody has written "tone filter" above the three resistors and three caps just before pin 19 on the percussion board. Could this be right and how to modify it? In addition there are some small pcb's behind the Brilliance tab that are connected to the percussion board pin 13(figure 5-17 console sheet 1, parts R3xx, C3xx,d3xx, etc.) I noticed that R357 affects the percussion length but could the rest affect the percussion sound?</P>


There are also a few resistors and two caps behind the vol soft and brilliance tabs. Can I try different values there without damage?</P>


Are there any filters on pedal divider and filter board that could be removed to add click to the pedal sound?</P>


Is the world round? Is there extraterrestrial life? What's the time? etc. So many questions, so little knowledge... [;)]</P>


Cheers,</P>


Jussi</P>

[tonewheel1966]

Re: Keyclick filters in T-series



Get Brendon to send you the recipe for "proper" vibrato chorus. We added a tiny transformer above the chorustab on our T serieseses and it sounds just like a B3/C3 vib chorus 3 setting. Everytime I play that setting on my C3 it reminds me of that little T series with the vib mod!</P>


Cheers</P>


Dave</P>

[Brendon Wright]

Re: Keyclick filters in T-series



I'm a bit late starting work, so I haven't got time to fill you in on details...</p>

But Jussi, we are SO in luck!!!!!</p>

I've just been working on a keyclick formula as we speak, it's still got some kinks in it though... it gives a dang fat click in the bass but it's distorting on the high notes.</p>

It involves bypassing the input resistors on the recovery amp with a capacitor, or "bright cap" as they call it in the guitar realm, then adding a big inductor to cut off the higher click frequencies... so essentially it's a 6dB band pass filter boosting the click in the midrange. Then I'm mulling over resistor values so that the click is stronger than the sustained sound but not distorting the high notes. I found a 1uF cap seems to be good, so far I'm using either a 7.5 or 9mH inductor... but.... </p>

I'll be back later on!</p>

Maybe some of the gurus from the electronics section might have pity on us and give us THE answer!</p>

-Brendoon</p>

[Brendon Wright]

Re: Keyclick filters in T-series



actually, on the schematics of the stuff behind the tab panel, afraid its all only on the main big ole logic diagram of the whole organ and you have to fish out the bits you need. Have you got that diagram?</p>

I also have photos of behind the T500 tabs and Dave sent me photos of his T200 behind the tabs, if I recall...</p>

Back later.</p>

-B</p>

[Brendon Wright]

Re: Keyclick filters in T-series



Jussi, Schematic from page 5-16 of the manual is the best you can get, I'm afraid:</p>

<span class="Apple-style-span" style="font-family: Arial; font-size: 10px; color: rgb(102, 102, 102); white-space: pre; "></span></p>

<span class="Apple-style-span" style="font-family: Arial; font-size: 10px; color: rgb(102, 102, 102); white-space: pre; "></span>I've decided on values and have pics on my new click filter, but I head out shortly, I'll post it if I have time but aren't certain if there's time...</p>

-Brendoon</p>

[Brendon Wright]

Re: Keyclick filters in T-series



Oh, yeah, and Kon Zissis's Chorus Vibrato Mod....</p>

I'll have to get onto that too!</p>

</p>

[tonewheel1966]

Re: Keyclick filters in T-series

[quote user="Brendon Wright"]


Oh, yeah, and Kon Zissis's Chorus Vibrato Mod....</P>


I'll have to get onto that too!</P>
<P mce_keep="true">[/quote]</P>
<P mce_keep="true">You've done that one [8-)]</P>
<P mce_keep="true">YOU sent me the transformer [:D]</P>

[T-200]

Re: Keyclick filters in T-series



Thanks Brendon,</P>


I have this one, but there are differences with my T. I have to figure out the lower vibrato on switch connection and related components. There are also differences because I have only brush and cymbal sound with volume, not the entire rhythm section.</P>


Sorry to answer this late, but I am very busy with my job and can't be here everyday. I really would like to here about your click filter mods and try them. I have a few ideas too, but I want to try them first...</P>


I bought some short alligator clip cables to try out some different values. Let's see what happens.</P>


Jussi</P>

[Brendon Wright]

Re: Keyclick filters in T-series

[quote user="tonewheel1966"][quote user="Brendon Wright"]

Oh, yeah, and Kon Zissis's Chorus Vibrato Mod....</p>

I'll have to get onto that too!</p><p mce_keep="true">[/quote]</p><p mce_keep="true">You've done that one [8-)]</p><p mce_keep="true">YOU sent me the transformer [:D]</p>

[/quote]</p>

Heh, sorry Dave, 'twas getting late. I meant get on to putting the details up here for Jussi!</p>

Hey Jussi!</p>

Do you have the T500 manual for comparison? I didn't realise your T had the rhythm unit like mine. The alligator clips are a grand idea. I'll post up these new click filter ideas today, plusa recording.</p>

Cheers!</p>

-Brendoon</p>

[T-200]

Re: Keyclick filters in T-series



Hi Brendon,</P>


It's getting late here again, but there's almost always time for a few lines.</P>


I found the brush/cymbal unit's specs inthe very sameschematics, I don't have any other rhythm unit. My T is model T-200 from '68 and the lower right panel has only brush switch/ cymbal switch/ lower manual vibrato on- off switch / brush&amp;cymbal volume pot. I have never seen a schematic with the vibrato lower switch. The percussion board is also different that I have seen anywhere. It's not the thyristor board or the board with IC's, but I got the percussion working OK with the external small pcb that I built as Carsten told.</P>


But no problem, I will figure it out when I have time. Let's change ideas with this thing, thoughI think that I have a lot less to offer than you do ...</P>


Kippis,</P>


Jussi</P>

[Brendon Wright]

Re: Keyclick filters in T-series



It was this time last year that I started into the Hammond modding, and with it, an introduction to electronics.</p>

My electronic understanding is still very shallow, but after a year of it you start to get ideas worth exploring.....</p>

Take tons of photos as you go, Jussi, and type down quick notes to refer to later.</p>

You can THEN write up a whole document on the thing.</p>

I find text without photos hard to picture.... but then I have all these photos I can't decipher because I can't always recall the EXACT details in retrospect. What were the component values I replaced? What were the originals? What were the results and what was the problem I was trying to fix?</p>

I've got to write a PDF on doing foldback and manual tapering at some stage, but the details are getting more hazy as time goes on...</p>

</p>

[Brendon Wright]

Re: Keyclick filters in T-series



Okay, I'm back!</P>


Firstly, here's the recording: Keyclick Trial.</P>


It's not perfect, but to my ears it was getting better. The ordinary electrolytic 1uF cap allowed a lot of tone thru, down to the punchy mids, but also a lot of the crispy trebles. The "bright cap" addedto the vibrato does that crisp treble thing.</P>


I think its almost too crisp. Check outhttp://www.drlonniesmith.com/for some samples of nice B3 click. So I added a whole string of inductor coils until it sounded about right. A multimeter which can measureinductance is a WAY brilliant tool if you're going to do this kind of thing, and make crossovers for your leslie and so on. 10mH was about right. A teeny transformer from a radio measured 11.75 mH which was close enough.</P>


The high octaves STILL distorted, probably because I'd already earlier done the manual tapering mod. This mod may even render the tapering mod unecessary!!!</P>


Anyway, I added two 82ohm resistors, totalling 164ohm. EVERYTHING was in series, so it was a VERY basic bit of electronics. The lot bridged the</P>


R856 resistor on the Recovery board.</P>


I'll post a PCB layout after these photos:</P>


Here's the experimental bench with various cannibalised inductor coils.</P>


</P>


Incidentally, for anyone who has a T with the drum machine under the lid, one mod I HIGHLY recommend is putting hinges on the back of the lid.</P>


I've had the lid propped up many times to fiddle inside the guff but had to leave the drums plugged in to keep the lower manual. </P>


Last week I removed the Drum Machine and have built the dummy plug. The Drums can still be plugged in as an external unit. </P>


Several times the lid fell off and broken things. One time it broke the lightswitch off. Another time it shattered the auto accompaniment board, which took quite a bit of glue and solder to resurrect!!</P>


Now I have the lid hinged, and its luvverly!</P>


Here's the circuit tidy on a perfboard (this one's for the upper, I haven't done the lower yet):</P>


</P>


And in place beside the recovery board, with a switch. </P>


</P>


I left a gap under the front lip of the lid, I'd like to mount the switches there. The gap is there primarily because of the depth of the lights. It's nice having the tungsten light spilling out over the keys though! Hugh from Hammond Zeni replaced his lights with an LED strip, bright, white, and takes no space. I looked upthe recipe forLED lighting strips online, and they're pretty straightforward. Maybe some other day when I decide the nice orange glow isn't for me!</P>


Now. The PCB:</P>
<P mce_keep="true"></P>
<P mce_keep="true"></P>
<P mce_keep="true"></P>
<P mce_keep="true"></P>

[Brendon Wright]

Re: Keyclick filters in T-series



Here's the recovery amp PCB:</P>


</P>


You can see the peppermint green bits are the bits Carsten Meyer had us change. Eg, for the upper manual, C856 is designed to give us more bass and body.</P>


Kon Zissis pointed out R856 as the Resistor to replace if I wanted to boost my upper manual volume. Here's the email here, which included the idea for this mod, adding a bright cap: http://organforum.com/forums/permali...ead.aspx#93760this was an additional email to the one which contained Kon's original T300 mods, and answered a few questions I had after his and Carsten's mods.</P>


I haven't discovered or asked what the other caps, resistors and trannies do on this board, but there's a lot of potential init if we DO find out, think of the mods.....</P>


You can see, anyhow, which bits are specific to the upper and lower buss amps.</P>


I'm going to ask expiano to see if he can explain things to us as a bunch of willing students!</P>


Ciao!</P>


-Brendon</P>

[T-200]

Re: Keyclick filters in T-series



I salute you Brendon The Great,</P>


I listened to your keyclick samples and that is just what I was looking for. Iseem to have started from the wrong end, the output of the rec.amp. I tried different caps with R870 and R871 with no results. I figured out that it would be the right place, because Carsten said in his mod that replacing C865 with a smaller cap would add more treble to the sound.</P>


Your way seems to be right, I can hear. In addition you can adjustupper and lowerseparately!</P>


I tried to figure out adding the keyclick to the bass pedals, too, but if I remember right, the pedal signal comes through the lower input of the rec.amp ... so if youmake the change in the lower stuff, it should also affect the pedal sounds, right?</P>


One thing I'd like to try is to balance the sound between nonvib and vib sounds. Do you think that changing the vibrato input parts in the rec.amp (R859 or R866, I don't know which one is the input resistor :-) would affect the soundwhen the vibrato is on?</P>


I'm on the road againand will be back homenext week, so I just have to keep this theoretical... and study these schematics in the hotel rooms.</P>


Dasvidanya, (as ournext door neighbours, the Russians, say [:P]</P>


Jussi</P>

[Guest]

Re: Keyclick filters in T-series



Brendon</p>

Saw your Bat signal, and I'm here to delight and enlighten...</p>

Slooshy well, my faithful droogies, as your humble narrator takes you down the garden path of devil-ish mods and... </p>

Wot!!?, harrumph! Stop that now, you hear!? It's silly. Now, get on wiv it, and none of this funniness!</p>

OK guys, take a big breath and hang on to your hats.</p>

I'm going to refer to the T500 recovery amp schematics from the Cap'ns website. It seems to match the photo posted on this thread. This will be a very light examination of the circuitry, with some very basic concepts explained. Those of you who are technically inclined, I can suggest you load the free LT-Spice simulator from Linear Technology onto your PC, enter the circuitry as you see it and run some AC analysis. Of particular interest will be the frequency response of the main recovery amp, Q854, who effectively shapes the spectrum according to the network in its feedback loop.

So let's start with the two manual recovery amps. They are identical so we only need to look at one, Q851, the lower manual amp. This is a very basic gain stage with collector fed DC bias, a topology found in preamp stages of many guitar amps and consumer audio gear. The two primary elements of the amplifier are its AC signal gain, and its DC operating point. I'll examine both in the following.

Lets consider the DC first. A key point to remember about the transistor (besides its current gain) is that it actively tries to maintain a constant voltage difference between the base and emitter, about 0.6-Volts for a silicon transistor. In this amp, the DC bias source that sets Q851 to its correct operating point is R851. This high value resistor trickles a tiny current from the collector into the base, since the collector is a few volts DC positive of the base. This is a self-stabilizing bias mechanism which depends on negative feedback to keep the transistor on a particular region of its transfer curve (explained later). To understand this, imagine for a moment what happens when the base voltage tries to move higher. The transistor reacts to maintain the base-emitter voltage constant, so it must raise the emitter voltage by the same amount. To do this, it pulls more current from its collector to increase the voltage on the emitter resistor. The higher collector current produces a larger voltage drop across the collector load resistor, which causes the collector voltage to move lower. Since the base bias resistor is fed from the collector, the lower voltage on the collector reduces slightly the base bias current, lowering the base voltage. This counteracts the transistor trying to increase the emitter voltage, and thus the control loop is closed. By picking the right value for R851 for a given current gain specification of the transistor, The circuit can be set to operate at a nominal DC point that will vary only somewhat over temperature and transistor production gain variations. Normally, to get maximum signal swing and variation immunity out of a stage like this, the bias is chosen to set the collector voltage halfway between the DC power rail voltage and the emitter idle voltage, at room temperature with a transistor that exhibits nominal specified DC gain. The transfer curve of the transistor (its gain versus the instantaneous voltages and currents) is fairly linear at this point, so as long as the output voltage swing stays under about 1V, very little distortion will result. As the stage is overdriven, the peaks in signal swing on the collector push out beyond the linear part of the curve and distortion starts to come in as the waveform peaks get flattened. Deliberate distortion can be introduced at normal signal levels, rather than overdrive levels, by shifting the bias point up or down (playing with the bias resistor R851 value). Typically, this results in even harmonic distortion (the nice sounding kind) due to the asymmetrical clipping that is introduced. Since there is some signal amplitude loss when this is done, the overall AC gain of the stage has to be raised a bit to compensate. Note that R851 has negligible effect on AC gain (to be explained next) since its over 100 times larger than R852.

Now for the AC. AC gain of the amp is primarily determined by the ratio of collector load impedance (resistance) R852 to the emitter resistor R853. Ignoring for the moment the effect of external loads present on the outbound end of capacitor C853, this can be understood in a very simplified but intuitive manner. A transistor is a current amplifying device; a current into the base lead is multiplied by the gain of the transistor (typically 70-150 times). This multiplied current flows between the collector and emitter. The emitter resistor thus sees a current through it that equals (base current) + (base current times device gain). The base current in the emitter can be ignored due to its very small size compared to the multiplied current coming from the collector. When the base is driven with an AC signal, very nearly the same AC signal appears at the emitter per the key point in the DC explanation, and the transistor draws as much current as it needs to from its collector to ensure this. Using the values from the T500 schematic, the 6mV (.006-Volts) applied to the base through C851 ultimately appears as a 6mV variation across R853. A 6mV variation across 680-ohms results in a variation in current through the resistor of 8.8-microamps (uA), per Ohms law. Now, if the gain of the transistor is 100, 99% of that 8.8uA current variation in the emitter is drawn from the collector, and only 1% from the base, so the collector current is for all intents equal to the emitter current. The collector can be fed from a moderately high voltage (20 to 40-Volts typically), through a much higher value resistor than the one on the emitter. This is where the voltage amplification occurs; the 8.8uA emitter current varation appears at the collector regardless of the collector resistor value. In this circuit, the collector resistor R852 is 10K (10000)-ohms, so, using Ohms law again, the 8.8uA variation produces an 88mV voltage swing on the collector. Since we started with 6mV input on the base, the circuit boosted the AC signal level by a factor of almost 15 times. The signal on the collector is inverted, due to the same mechanism as explained for the DC bias loop. An upward swing of the input signal waveform causes the transistor to draw more current from its collector, and thus the voltage on the collector will swing down, scaled by the deliberate gain factor of the stage.

Now I'll talk about the mods on the manual amps. Capacitors C851 and C853 couple the AC signal into and out of these amps. A transistor needs certain DC voltages on its pins to operate correctly as explained earlier, so the capacitors allow AC to be driven in or pulled out without disturbing the DC bias. A side effect of this is that capacitors form a high-pass frequency response (bass rolloff) that's a function of the capacitor value and the total impedance (load + source) they connect between. A single R-C section creates a single-order rolloff, that is 6dB per octave, set at a frequency (pole) that can be calculated by the simple formula Fc = 1/(2*Pi*R*C). Since there are two independent R-C networks (input and output), the effect of both is summed across the spectrum, and the low frequency rolloff eventually becomes 12dB per octave. For full audio band amplifiers such as these, the low frequency cutoff must extend below the frequency of the lowest note expected. As a general rule, it should be at 1/2 the lowest frequency so as not to affect the in-band signal level. On the input side, the impedance the transistor presents in the desired audio band is well over 10K ohms, which forms a high-pass pole with C851 at around 2.3Hz. This is fine. On the output, however, the capacitor is 1/30th the value, so at the same impedance, the pole is at over 60Hz. It could actually be lower but I don't have the full schematic of where that signal feeds and what load it drives. 60Hz is around two octaves down from middle C, so anything around that octave or lower will start sounding thin since the signal is already down 3dB (1/2 power) at the pole frequency. It's also possible, with the age of the organ, that the capacitor values have eroded somewhat and the pole has crept up in frequency. However, it's still rather odd that Hammond set the input pole at such a low frequency but left the output so thin on bass. The small cap from base to emitter provides a high frequency rolloff, probably well above the audio band so its effect can be ignored. That's likely why the Kon mod didn't include changing those. They should be left as is to ensure high frequency stability of the circuit, since the transistors probably have Ft's up in the 100MHz region. The gain of the circuit can be increased by lowering the emitter resistor value as suggested by Kon. In order not to disturb the DC bias, however, a resistor in series with a cap should be added in parallel with the existing 680-ohm. The value of resistor to add depends on the gain increase you need. The transistor should be able to deliver more than double the gain it presently does, so you could go as low as 680-ohms with the added resistor. Chances are, however, that you only need a modest gain increase. A 6.8K resistor will give you a 10% gain boost, a 3.3K around 20%, 1.6K around 40%. If you use a 5K trimpot with a series 1K (plus the cap), you can get variable gain boost from about 15% up to about 30% or so. This will allow you to fine adjust the level so they match between manuals very closely. The series cap forms another high-pass pole, and the value should be chosen to set the pole around 10Hz or so. Use the formula with the total series resistance value; it should be a good approximation.

Finally, the recovery amp, Q854, and its follower buffer, Q855. This is a key stage for sound color since it contains a deliberate spectrum shaping network. The manual amps are essentially flat response through the passband and should remain so. It's best to do all the frequency shaping at one stage, and it makes sense to have it here at the final recovery amp, right before the final power amp. The feedback network C860 to C862, and R861 to R863, form a phase-shift network that produces more than just the simple high frequency rolloff (single pole) of a single Miller capacitor. I ran a quick LT-Spice simulation (much quicker than figuring it out on paper). The whole amp stage exhibits about 17dB gain in the passband (100Hz to 3.5KHz), and imparts around a 6dB boost bump at around 3.5KHz, just before a 24dB/octave rolloff starting at 4KHz. The R-C networks on the base of the buffer transistor Q855 (R867 to R869, C863 to C865) work in concert with the feedback network on Q854, almost doubling the rolloff rate above 4KHz, and also shifting the bump peak to around 3.5KHz while adding about 1dB of more peak height. The keyclick primary energy spectrum is above 3KHz, so the effect of this circuit will be quite pronounced on the click sound. Moving the peak from lower frequency to higher frequency, the click sound will progress from "pup" to "pop" to "pip" and finally to "tick". The depth of the click resonance will be determined by the peak height and narrowness in relation to the passband gain. A high peak will impart a definite note to the click, like striking a small hard wooden peg, the resonance note being the peak frequency. To play with this circuit's behaviour without changing too much at once, you can do the following: scale the value of the six capacitors C860 to C865 up and down to move the frequency of the peak and rolloff up and down without changing the overall shape of the rolloff and peak height. Make them smaller to move the frequency upwards, and larger to move the frequency downwards. Changing the peak to passband gain ratio is a bit trickier, but will probably produce more sonically interesting variations. The peak's "Q" is determined by the set gain of the Q854 stage. increasing the gain of Q854 (by reducing the value of R866) not only shifts the overall passband gain, but dramatically changes the peak to passband ratio. All this, without moving the location of the peak or rolloff significantly. Referring back to the manual amp section, the way to change gain is to add a series R-C in parallel with the emitter resistor in order to preserve the DC bias point. In the case of the manual amp, we wanted a wideband capability, so the capacitor will be deliberately large to extend response down to bass frequencies. In this circuit, however, we want to maintain the passband gain largely where it is (around 17dB) and just affect the peak height. This means that the series R-C will have to be tuned to have an Fc above 1KHz. I've added one in the simulation, .22uF in series with 270-ohms. The passband gain remains relatively flat to about 2KHz, then the peak rises a dramatic 10dB before the rolloff slope. This should produce a very pronounced tonality to the keyclick, and the peak level can be adjusted by changing the resistor and capacitor values, but keeping the R-C product the same (decrease C value by the same factor that you increase the R). As you increase the R (and decrease C), the peak will drop back toward the level that it currently is.

So, to sum up, I've provided a highly simplified view of the stages you will need to modify, along with some simple rules to ensure the changes can be done in measured steps and without interacting too much. There are ways of introducing inductors here, but I've tried to keep the parameters simple so that minor changes in existing values, and the strategic addition of R-C networks, should achieve a usable level of sound modification. Of course, there is much much more to the design of these stages than what I've explained in these brief paragraphs, but you are now armed with enough background to experiment without too much frustration.
</p>

Enjoy,</p>


exp
</p>

</p>