gain staging and attenuation

The JC3 specs 48dB gain in MM mode. That’s a lot. Though it’s unclear whether this is affected by choice of outputs RCA vs. XLR. Ortofon 2M are also known for having hot oputput levels, > 5mV. This is going to result in an unusually high output level for a vinyl source. That explains your observations. You’d be better off with an MM cart in the 2.5mV range. 

Adding extra attenuators downstream means you’re generating too much extra gain you didn’t need. It’s "suboptimal", though most of our systems do it to some degree. On the question of where to place fixed attenuators (like Rothwells) - you might notice slightly different levels of attenuation based on where you insert them, due to the different input impedance of preamp versus power amp. You can sometimes get a lower noise floor by placing the attenuators between preamp and power - particularly if it’s a tube preamp w/ high gain and your speakers are sensitive. In that case, the attenuator helps de-emphasize the active preamp’s noise floor. But in most cases, it probably doesn’t matter much where you choose to put the attenuators. 

The situation will likely be different if you switch to an MC cart. JC3’s gain spec for MC mode is 64dB, which is more modest than its MM gain. This would pair perfectly to cartridges ~ 0.4mV. 

This is also a good illustration of why multiple gain options is a very good feature for phono stages. With one fixed gain (per MM/MC mode) you’re limited in your "optimal" carrtidge matchings. 

For a passive preamp insertion (Goldpoint, Khozmo etc) you’re going to have to be careful about imedance matching. A passive volume control serves simultaneously as the load impedance for the upstream and the source impedance for the downstream. First step is to list out the input and output impedances of your components. That might help choose where to put it. Then you can select the correct impedance for the passive control (usually 5K - 10K). You want to aim for at least a 10x imedpance ratio for each source-to-load interface (more than 10x is better). 

Example:

• Source component output impedance: 500 ohms
• Load component input impedance: 50,000 ohms
• Passive attenutor of 5,000 ohms value: On the source size: 5,000 / 500 = 10x (Good). On the load size: 50,000 / 5,000 = 10x (Good)  
• This is cutting it close to the 10x rule of thumb on boths sides of the passive, but it will work fine.
• Tube preamps in particular might not like driving lower impedance loads, like a 5K - 10K passive, even if you hit the 10x ratio.

Hmmm. Probably not what you would like to do, but I would just use the preamp...it has appropriate switching & gain options. If you don't like the sound, look for a preamp you do like the sound of. This is exactly what they were made for.

Based on mulvelings detailed instructions (thanks!) there are not many ways to do this. Passive pre has to go before the ARC preamp because the input impedance of the power amp is about 22K, and the passive pre is 10K (or maybe even 25K). From reading, my understanding is that the passive pre will have the same output and input impedance, and that the impedance is barely affected by whatever volume setting you use on that pre. Should the attenuator also be treated as a separate device with input and output impedance the same?

Kr4 - Thanks for the helpful pointer. I guess it is best to prioritize impedance matching over putting gain first, though effects on noise make sense. 

Audio_guy - at acceptable volumes the ARC is at about 7 or 8 o’clock while  on its lowest of 3 gain settings. Frankly, it is pretty old and seems to sound better when used in bypass mode  New pre would come in handy.   

Like was mentioned before, the Parasound phono stages have a lot of gain, comparatively.  Are you sure it is set to MM on both channels and not MC, by accident, giving you an insane amount of gain?

Just a final note to write up the solution here. Thanks for everyone’s help. I was using the variable attenuation of a passive pre to adjust volume because I wanted to leave the ARC at unity gain. I thought it sounded markedly worse with volume controls etc engaged. I did want to have that ARC in there to buffer the passive pre.
 

After much experimentation I ended up getting rid of the passive pre and simply using the ARC volume control, but putting a Rothwell 15 dB attenuator in the input of the ARC.  This enabled me to use the ARC volume at around 12 o’clock with its gain set at low. Certainly sounded better than using the volume at 7-8 o’clock. Noticeable improvement in the richness of the sound and the soundstage. Still not sure if the “low medium high” knob on the ARC alters gain and the volume knob is attenuation of that gain, but this works. It clearly sounded better to use the attenuation at 12-3 oclock (rather than 7-9 o’clock) even if that meant using a lower “gain” setting. Shout out to the Rothwell attenuators. No noticeable effect on sound quality at this position in the chain and enabled me to use optimal settings on the pre, just as Rothwell describes.