Simplistic NJFET RIAA

Member
Joined 2009
Paid Member
Copydex is a liquid.
It seeps in around the debris and surrounds it. That makes it easy to remove the debris, as the skin of partially set rubber film is peeled out. It does no harm to leave it a bit longer until all the film has set to clear, about 5 to 15minutes if thin enough and approaching 30minutes if you apply a big glob.
Blue tac is a solid and will not flow around (no capillary action). It also does not peel well and tends to leave some blue tac behind which will be difficult to ensure 100% of the blue tac. You still have the problem of the debris lying in the gaps that the blue tac has not surrounded.

Please AndrewT could you buy for me the Copydex? I can't buy here in Spain.

Infact, I dont think the cartridge from Merlin is in big danger also it would be worth to remove this parts.

It is important to be aware of this possible problem, specially when
removing the transparent protection sheet.

Digicams can be very revealing for such things, isnt it?

Franz

Still not removed the transparent protection, it's necessary to remove?
 
Member
Joined 2009
Paid Member
So you clearly hear differences just by adding 18p in C3 :)

Yes I hear differences adding 18pF in C3 also I hear differences adding 4.7pF in C2, now sounds better with 18pF more but seems there is more room to add other 15 or 18pF I will keep informed

Another important remark.... Never work with iron tools near the cartridge because MC´s have very powefull magnets that can pull a screwdriver with tremendous force.

Very very good advice, it's a easy way to damage the cartridge.
 
Last edited:
Please AndrewT could you buy for me the Copydex? I can't buy here in Spain.
I have not tried it, but I wonder if anyone has experience of using rubber solution as in repairing punctures in bicycle tyre tubes could be used instead.
This has more STICK than Copydex and may have more aggressive chemicals that might have an effect on plastic parts.

Codydex is relatively non aggressive and has low tac.

What are the rules about posting a container of liquid in the EC?
 
Member
Joined 2009
Paid Member
That source on Ebay notes worldwide shipping. Copydex 125ml. This cheaper post cost one excludes only Canada and Italy. Copydex.

Hey Nick thanks for the links.

Hi Merlin

If you go away from 47n + 10pF for C2 you will not have a correct 500.5Hz mid freq point for the riaa.

Remenber that you have R4 = 6764 so 6764 x 0.04701 = 318uS (500.5Hz)

Possible I have a little bit more than 47nF & sounds better than adding more capacitance.

Definitive now C3 needs more capacitance, I will try tomorrow.
 
There is guessing and there is trying empirically.
an error in one time constant cannot be compensated by introducing errors whether guesstimated or empirically derived in the other time constants.
The erroneous time constant must be corrected and then all the other time constants must be maintained by rescaling all the component values to suit that one erroneous capacitor.

In other words,
If you know you have one erroneous capacitor then the solution is to correct only that one capacitor.

Or more simply:
If the capacitor is too big, throw it away and fit the correct value capacitor. There really is no alternative !
 
an error in one time constant cannot be compensated by introducing errors whether guesstimated or empirically derived in the other time constants.
The erroneous time constant must be corrected and then all the other time constants must be maintained by rescaling all the component values to suit that one erroneous capacitor.

In other words,
If you know you have one erroneous capacitor then the solution is to correct only that one capacitor.

Or more simply:
If the capacitor is too big, throw it away and fit the correct value capacitor. There really is no alternative !

An error in capacitance measurement by DVM, especially with capacitance values discussed here, can and should be compensated.
Stray capacitance and inductance in a circuit cannot be measure directly by most hobbyists, and they need to be compensated, especially in the RIIA equalization section.
 
An error in capacitance measurement by DVM, especially with capacitance values discussed here, can and should be compensated.
I can understand compensating a slightly low value capacitor by paralleling a second to bring it up to design value.
How do you compensate a slightly high value? Certainly not by trial and error component changes in all the other time constants.

I would propose that the slightly high Cap value be compensated by matching it with a slightly low value resistor to give the exact RC time constant that the EQ requires and then rescaling all the other R&C values to bring them back into line with the compensated one.
Do you see another way?
 
Last edited:
That is what we have been doing
What have any of you been doing?

What I am reading and supposed to believe is that small errors can be compensated by adjusting the other RC time constants by willy nilly swapping/paralleling other component values. This can never achieve the correct RIAA.

If one capacitor is too big, then the whole RIAA must be recalculated to match that one cap and then assembled and listened to.
If one cap is too big, one must correct that error first. This is NOT achieved by leaving in the same resistor value. That guarantees the wrong time constant. and other swaps can never eliminate that RC error!!!!!!!
 
Member
Joined 2006
Paid Member
Merlin measured all the components and we reached the following good values:
R4 = 6764 so 6764 x 0.04701 = 318uS (500.5Hz)

What I am saying is that we can not be tottaly sure these values are correct because we depend on the DMM readings.

The cap measures 47n so I suggested adding a 10pF // in order to reach 47.01nF.
 
Last edited:
Are you accepting that resistor tolerance and capacitor tolerance must both be to the same standard to achieve a RIAA EQ to within a set tolerance?

Then measure your components to the best tolerance you can manage. That will almost certainly require some other method of measuring capacitors than using a DMM with cap function. Yes, you can select two 47nF caps to be near enough equal to the resolution of the meter. But you cannot measure within the same tolerance, cap values that are over a few decades of difference in capacitance values.
Herein lies the problem.
Various methods can be used to remove or reduce errors and tolerances in measuring resistors that vary over a few decades of resistance value. But our Cap Function in a DMM cannot do that. It's the "method" that will determine the capacitance selection tolerances.

I can give you an example.
A Sallen & Key equal component value 2pole active filter uses the same value of components around the circuit and between channels and between multi-stages if required.
Adopt a unity gain Sallen & Key 2pole active filter and now some components are equal. But two critical component values not only have to be the same between channels and stages, but they must also be in an exact 1:2 or 2:1 ratio as the other selected tight tolerance components.

The "easy" way to achieve that is to use two equal value components tied in series (resistors) or tied in parallel (capacitors) to get that critical 1:2 ratio and for identical values between channels and stages. This matching of component values determines the tolerance of the EQ achieved. Unity gain has an "easy" solution, MFB requires a different technique if the same tight tolerance needs to be achieved.

The worst tolerance components will determine what can be achieved with RIAA EQ accuracy. Changing the wrong components cannot correct the error and improve the RIAA EQ tolerance.
 
Last edited:
I can give you an example.
A Sallen & Key equal component value 2pole active filter uses the same value of components around the circuit and between channels and between multi-stages if required.
Adopt a unity gain Sallen & Key 2pole active filter and now some components are equal. But two critical component values not only have to be the same between channels and stages, but they must also be in an exact 1:2 or 2:1 ratio as the other selected tight tolerance components.

And how would you take into account the stray capacitance and inductance of that Sallen & Key filter?

You see, dealings with such small values of capacitance, accurate measurements necessitate a professional and very expansive measurement gear.