I'm doing some improvements to an ARC SP-4A. They load the RIAA input with 100k to ground, then go through a 150 uF coupling cap, followed by another 100k to ground. The AC input impedance is thus about 50k down to some very low frequency. They need the coupling cap because the amplifier module can have up to a volt or so present on the input.
My question is, why such a high value of coupling cap into a 100k load? I'd think just a few uF would be plenty, and it could be a film, rather than an electrolytic. ARC even used a tantalum, IMO not the best choice.
Is there any interaction with cartridge inductance or something I'm not considering that justifies the huge value?
My question is, why such a high value of coupling cap into a 100k load? I'd think just a few uF would be plenty, and it could be a film, rather than an electrolytic. ARC even used a tantalum, IMO not the best choice.
Is there any interaction with cartridge inductance or something I'm not considering that justifies the huge value?
Last edited:
The thinking is to have a very low freq -3 dB point and the belief that over-sized ecaps have lower distortion.
You did not state what the voltage rating of that ecap is? I would think something like 10-25V would have been used
Yes a film cap would be a better choice, something like a 1uF/50-63V film. This would give you a -3dB of 3Hz. Polyprop is the better dielectric over mylar, polyester although more expensive & larger in size.Wima makes some nice parts.
You did not state what the voltage rating of that ecap is? I would think something like 10-25V would have been used
Yes a film cap would be a better choice, something like a 1uF/50-63V film. This would give you a -3dB of 3Hz. Polyprop is the better dielectric over mylar, polyester although more expensive & larger in size.Wima makes some nice parts.
Yup, that's definitely part of the plan. Right now I'm contemplating the RIAA values used. Part of the problem is zero knowledge of what's inside the ARC AM-6 potted module. If I treat it as an opamp with an extra parallel feedback path, LTSpice gives me pretty accurate results. The measured frequency response using an inverse network is quite good, probably better than .25 dB, but the impulse response to a 300 Hz square wave shows some overshoot that I think makes it a bit bright. My DIY LME preamp doesn't do that and I think it sounds better. I'm just not ready to replace the ARC circuit with my own without giving it a fair chance via some tweaks to the RIAA feedback values.
If one works out the C R point with 150uF and 100K it is substantial. Maybe if they wanted to approach DC they should have designed the pre amp accordingly using a balanced input that works at 1.5mV from the moving coil pickup.
It, at this point. would be useful to take a look at the frequency response of the recording device used to produce the record in the first place. I remember EMI stating the bottom end frequency to be in the order of 27HZ so why go down to less than 15HZ, all that will be heard is road noise or a noisy rumble from tired turntable bearings. The VLF will only have to be filtered out anyway!
It, at this point. would be useful to take a look at the frequency response of the recording device used to produce the record in the first place. I remember EMI stating the bottom end frequency to be in the order of 27HZ so why go down to less than 15HZ, all that will be heard is road noise or a noisy rumble from tired turntable bearings. The VLF will only have to be filtered out anyway!
Are you checking the impulse response with the inverse network in place, or without it?
Network in place. It's the one from the Hagtech paper. My DIY preamp shows both flat frequency response and good impulse response, which I'd expect. I'm a bit baffled why the ARC also seems to be reasonably flat, yet has a bit of overshoot and recovery problem, unless it takes less response error than I would have expected to cause an overshoot problem. LTSpice says I need to alter the ARC values a bit to improve both- I assume they didn't have LTSpice back in 1976, and so might be forgiven for not getting it quite perfect.
I'd trust LTSpice on this one, but only after making 100% sure your inverse RIAA network is correct. The spread below will step you through the Lipshitz equations (active equalization only) and does a pretty good job. You can get to plus minus 0.1 dB with a bit of tweaking.
Lipshitz remarked in his paper that a lot of phono amps he tested (this would be up to 1980) had equalization problems - and you still see that today sometimes.
http://hifisonix.com/wordpress/wp-content/uploads/2010/10/RIAA_Calculator5.xls
Lipshitz remarked in his paper that a lot of phono amps he tested (this would be up to 1980) had equalization problems - and you still see that today sometimes.
http://hifisonix.com/wordpress/wp-content/uploads/2010/10/RIAA_Calculator5.xls
This module could use feedback bias stabilization, an example circuit:
2n2222a phono preamp
In this circuit, a large input cap may be needed for low frequency stability, check for this if you reduce the value. But 150µF seems large.
2n2222a phono preamp
In this circuit, a large input cap may be needed for low frequency stability, check for this if you reduce the value. But 150µF seems large.
- Status
- Not open for further replies.