I have been asked why there is a need to match for transconductance, when we already matched tightly for Idss.
The attached graph are two 2SJ74BLs with identical Idss, actually measured with a curve tracer.
As you can see, real devices do not obey theoretical equations.
To match for transcoductance for F5 without a curve tracer, the following procedures are sufficient :
2SK170BL -- measure Id with 0R, 11R, 22R degeneration
2SJ74BL -- measure Id with 5.1R, 16R, 27R degeneration
Devices (NNPP) with identical values for all 3 columns are best matched for the F5X.
Patrick
PS These procedures will be described in more detail by the test team eventually
.
The attached graph are two 2SJ74BLs with identical Idss, actually measured with a curve tracer.
As you can see, real devices do not obey theoretical equations.
To match for transcoductance for F5 without a curve tracer, the following procedures are sufficient :
2SK170BL -- measure Id with 0R, 11R, 22R degeneration
2SJ74BL -- measure Id with 5.1R, 16R, 27R degeneration
Devices (NNPP) with identical values for all 3 columns are best matched for the F5X.
Patrick
PS These procedures will be described in more detail by the test team eventually
.
Attachments
post403:
for an ammeter use a 240r 0.1% resistor and a 200mVdc DVM with a 0.1% scale accuracy. This is likely to be 100 times more accurate than using an ammeter.
Or settle for 10 times more accuracy using 1% devices.
for an ammeter use a 240r 0.1% resistor and a 200mVdc DVM with a 0.1% scale accuracy. This is likely to be 100 times more accurate than using an ammeter.
Or settle for 10 times more accuracy using 1% devices.
the standard test is Vds=10V and Tc=25degC.
If either of those vary, then your results will be different.
A battery may be stable enough if it is fresh, but be very wary of drooping voltage if the test session is extended.
If either of those vary, then your results will be different.
A battery may be stable enough if it is fresh, but be very wary of drooping voltage if the test session is extended.
I use regulated supply. So the supply was not an issue.
Temperature is. Breathing on it will swing the Id by as much as 50uA.
So although my ammeter can measure to 1uA, I never quote better match than 20uA.
If you connect the JFET and just watch it from 2m distance over 5 minutes, it will still be drifting to within 10-20uA.
Patrick
Temperature is. Breathing on it will swing the Id by as much as 50uA.
So although my ammeter can measure to 1uA, I never quote better match than 20uA.
If you connect the JFET and just watch it from 2m distance over 5 minutes, it will still be drifting to within 10-20uA.
Patrick
Quick (newbie) question regarding gain. From a post in the old thread, it sounds like the gain of the F5X will be the same as the F5 (~5.6 or 15dB). If I am running the amp with a simple buffer with no gain for a preamp, how many Volts (Vrms) does my source need to output to push the amp to its maximum power output?
Careful !
balanced lines can be specified in different ways.
If you feed in one half of the balanced impedance signal as a 1Vac and the other half as an out of phase 1Vac then the F5x will amplify each half to give ~6Vac from each output.
The actual output is 12Vac due to the out of phase feature of the balanced amplifier.
That is not a gain of 12times, it is a gain of 6times.
balanced lines can be specified in different ways.
If you feed in one half of the balanced impedance signal as a 1Vac and the other half as an out of phase 1Vac then the F5x will amplify each half to give ~6Vac from each output.
The actual output is 12Vac due to the out of phase feature of the balanced amplifier.
That is not a gain of 12times, it is a gain of 6times.
Thanks Andrew.
So if I am understanding this correctly, it sounds like my buffalo dac with an balanced output of 4Vrms (2Vrms +/-) should be enough to run the amp at max power of about 50w into 6ohms (class A).
So if I am understanding this correctly, it sounds like my buffalo dac with an balanced output of 4Vrms (2Vrms +/-) should be enough to run the amp at max power of about 50w into 6ohms (class A).
Standard CD output should be 2Vrms single ended, or 4Vrms balanced.
The F5X has a gain of 6. So 4Vrms x 6 = 24 Vrms, or about 34V peak to peak.
The F5X is designed to work at say 2V below rail, so 13V output on one half, or 26V peak to peak.
That would not leave you much for volume attenuation.
But then again you might not need maximum output for listening.
Patrick
The F5X has a gain of 6. So 4Vrms x 6 = 24 Vrms, or about 34V peak to peak.
The F5X is designed to work at say 2V below rail, so 13V output on one half, or 26V peak to peak.
That would not leave you much for volume attenuation.
But then again you might not need maximum output for listening.
Patrick
horio is probably using the sabres digital volume, why do you need the buffer mate? the Zout is low enough as it is with the D1 IV and if you arent adding any gain at all with the buffer why is it needed? i have never found the need to use anything between 9018 or 9012 and an amp, sabre's digital volume is near perfect imo, the output is hot enough
I plan to plug my Buffalo straight into the F5X and use the digital volume. Long term, I am going to throw a DCB1 in between for input switching and buffer purposes with my other sources (phono pre, another DAC and SACD player).
I just wanted to be sure I setup my Buffalo with sufficient Vrms out to use all the F5X power. =)
I just wanted to be sure I setup my Buffalo with sufficient Vrms out to use all the F5X power. =)
Spencer has shipped 40 quads matched 70/74 fets to me.
They are grouped in 4 current ranges:
Group K170 J74
1 6.90 6.95 8.10 8.15
2 6.95 7.00 8.15 8.20
3 7.00 7.05 8.20 8.25
4 7.05 7.10 8.25 8.30
He also did a transconductance comparison within the K170 and the J74 groups.
These quads are most likely the best suitable parts he could provide based on his new batch of 2000 J74's for the F5X. (J74 is likely close to becoming obsolete)
K70 is still available in big quantities.
I am happy to offer these for cost to all F5X builders, that have already ordered the cases.
They are grouped in 4 current ranges:
Group K170 J74
1 6.90 6.95 8.10 8.15
2 6.95 7.00 8.15 8.20
3 7.00 7.05 8.20 8.25
4 7.05 7.10 8.25 8.30
He also did a transconductance comparison within the K170 and the J74 groups.
These quads are most likely the best suitable parts he could provide based on his new batch of 2000 J74's for the F5X. (J74 is likely close to becoming obsolete)
K70 is still available in big quantities.
I am happy to offer these for cost to all F5X builders, that have already ordered the cases.
Here it comes another thing I had in mind: the ultimate flexibly pre-amp, to mix balanced and un-balanced source and amps.
D.
