Passive pre.... what's going on with this one?

[rabbitz]

I removed an active pre circuit for repair and left the remainder in so it could still be used. Shock, horror as it sounded very good indeed.

It's a bit different to a normal passive as the signal flow goes through several passives that were part of the active circuit after the pot.

CD Zout 100R > A20K pot > 2.2uF MKP cap > 68R in series with 100K to ground > power amp Zin 47K.

All the other passive pre's I've tried seemed to be a bit rolled off and lost some dynamics but doesn't seem to be the case here.

Is there benefits using passive components in conjunction with the pot in passive pre's or is this just an accident or an act of god?

[AndrewT]

Hi,
passive pre-amp can give very good performance if all the ancilliaries are correctly proportioned to allow that performance to get out.

Since you didn't design it this way you have just been lucky to stumble into a well matched set-up.

Now that you know it can be done, there may be a little optimising left to improve things still further.

Start with cable capacitance and poweramp input filter components.

Give us all the info.

[gootee]

Originally posted by rabbitz
I removed an active pre circuit for repair and left the remainder in so it could still be used. Shock, horror as it sounded very good indeed.

It's a bit different to a normal passive as the signal flow goes through several passives that were part of the active circuit after the pot.

CD Zout 100R > A20K pot > 2.2uF MKP cap > 68R in series with 100K to ground > power amp Zin 47K.

All the other passive pre's I've tried seemed to be a bit rolled off and lost some dynamics but doesn't seem to be the case here.

Is there benefits using passive components in conjunction with the pot in passive pre's or is this just an accident or an act of god?

This circuit, if I've interpreted your description correctly, is basically just a high-pass filter, with a response that rolls off at about 20 dB per decade, below its cutoff frequency.

I plopped the circuit into LTspice and ran some AC Analysis (frequency and phase response) simulations.

I used an ideal voltage source as the CD, with 100 Ohms in series with it. I then have the 20K pot between signal and ground, with the wiper sending the signal onward to the 2.2uF cap, after which there's a 68 Ohm series resistor, 100K to ground, and then 47K to ground to represent the power amp's Zin. For the 2.2uF MKP cap, I used an ESR (Equivalent Series Resistance) of .005 Ohms.

Looking at the "output", just after the 68 Ohm resistor:

With the pot set for maximum volume (.001 Ohm/20K), the high-pass cutoff frequency, f(-3.1dB), is about 2.8 Hz, and there's about -0.1dB of signal at the power amp input. Phase shift at 10 Hz is about 12.7 degrees, and is about 1.3 deg at 100 Hz. After adding 0.5pF of parasitic capacitance to each resistor, and 3 nH parasitic inductance to the 2.2uF MKP capacitor, the cutoff frequency was about 2.26 Hz, and phase shift at 10 and 100 Hz were about 12.45 deg and 1.28 deg. The frequency response now begins to fall, after about 30 MHz, but has a small peak at around 1.8 GHz.


With the pot set to its midrange, 10K/10K, the cutoff frequency (f(-10.35dB)) is about 1.96 Hz, and there's about -7.35 dB of signal at the power amp's input. Phase shift at 10 Hz is about 11 degrees, and is about 1.12 deg at 100 Hz. After adding 0.5pF of parasitic capacitance to each resistor, and 3 nH parasitic inductance to the 2.2uF MKP capacitor, the cutoff frequency was about 1.95 Hz, and phase shift at 10 and 100 Hz were about 11 deg and 1.12 deg. The frequency response now begins to fall, after about 1 MHz, but levels out and then has a peak of -5.9 dB at around 1.8 GHz, before continuing to fall.


With the pot set for low volume, 19K/1K, the cutoff frequency (f(-29.33dB)) is about 2.2 Hz, and there's about -26.33 dB of signal at the power amp's input. Phase shift at 10 Hz is about 12.4 degrees, and is about 1.26 deg at 100 Hz. After adding 0.5pF of parasitic capacitance to each resistor, and 3 nH parasitic inductance to the 2.2uF MKP capacitor, the cutoff frequency was about 2.2 Hz, and phase shift at 10 and 100 Hz were about 12.4 deg and 1.26 deg. The frequency response now begins to rise, after about 1 MHz, but levels out and then has a peak of -6.75 dB at around 1.8 GHz, before falling again.

- Tom Gootee

http://www.fullnet.com/~tomg/index.html

[Panelhead]

If your CD player is cap coupled, take out the 2.2 ufd capacitor. Also you can disconnect the 100K to ground. The 20K pot can be direct connected to the amplifier if desired.
A 10k log pot is a better match to an amplifier that is 47K. The 20K seems to be working okay for you.
If you have to leave in the 2.2 ufd capacitor, the 100K resistor to ground can still be disconnected.
Passive controls can work really well. But as said, the conditions have to be right. Some amplifiers are just not passive friendly. Same goes for sources.

George

[georgehifi]

Originally posted by rabbitz
I removed an active pre circuit for repair and left the remainder in so it could still be used. Shock, horror as it sounded very good indeed.

It's a bit different to a normal passive as the signal flow goes through several passives that were part of the active circuit after the pot.

CD Zout 100R > A20K pot > 2.2uF MKP cap > 68R in series with 100K to ground > power amp Zin 47K.

All the other passive pre's I've tried seemed to be a bit rolled off and lost some dynamics but doesn't seem to be the case here.

Is there benefits using passive components in conjunction with the pot in passive pre's or is this just an accident or an act of god?

With that setup your -3db @ 14hz, hardly what I would call good low frequency performance, besides this your just hearing what getting rid of unnecessary crap in the signal path sounds like, with an acceptable impedance match from cd to passive to amp.
Now get rid of the 2.2uf 68R and 100k and it should be even better.

Cheers George

[audio3000]

It also depends on the characteristics of your inter-connects.

I built a passive pre-amp several months ago and found that it lost too much high and brightness. I ended up have to make it an active zero gain pre-amp. The conclusion is that you need an active amp or a circuit with low impedance to drive inter-connects with high capacitive load properly.

[Conrad Hoffman]

As expected, these things work best with short runs of the lowest capacitance and good dielectric cable you can get. I have a buffer amp in my preamp, with a switch to take it out of the circuit. The extra gain is rarely necessary, and with short low cap cables, it sounds just fine. This is one instance when cables do make an audible difference. It's also important that the input impedance of the power amp is flat over the audio range, but fortunately, most are.

[georgehifi]

We had a cable shootout at one of our audio club meetings between about 20 different interconnects. 40 odd members.

All were measured for capacitance, and the general concessus was that the lower capacitance < 100pf per foot cables sounded the best, and it so happened that generally they were also the most expensive ones, except for the odd one.

This was using an active pre with very low output impedance. And then a passive pre ( Lightspeed Attenuator), the cd source was 50 ohms output. Power amp in was 100K

The same result was noted with both pre's, active or passive that it's better to have low (<100pf per foot) interconnects.

Cheers George

[rabbitz]

Thanks for your replies guys as I appreciate it.

Passive pre's are simple in concept but one of the hardest things to get right especially for the impedance challenged. I've read about the 100:1 rule on impedance matching but like most, I will not be able to achieve it.

The CD is a Sony CDP-X55ES which has a 100R resistor at the output. The pot is a 20K Alps remote RK27 series. The amp is a AKSA Lifeforce 55 and I've rechecked and the Zin is 42K and not 47K. The cables have a capacitance of 32pF/m so should work well.

I'll drag out the soldering iron this arvo and have a play and report back. I'll compare it to the SS pre that uses a 10K Alps black beauty in conjuction with the solid state section of an AKSA GK-1.

[AndrewT]

Originally posted by georgehifi
With that setup your -3db @ 14hz, ..........Now get rid of the 2.2uf 68R and 100k and it should be even better.

Hi George, did you lose a decimal place or could you explain this result?

Rabbitz,
we do need the cable lengths and power amp input capacitances to complete this analysis and find your optimum settings.