Hi Andrew
I have tried connecting a different CDP and its the same, as is the sound of the transformer when I short the RCAs
I have just rebuilt my LDR attenuator, and I have grounded it in the same way as my last. When either attenuators were connected, the amp hummed. But the previous didn't when connected to a chipamp. The chipamp is no longer assembled.
'What is the measured noise coming from the terminals'
How would I achieve this?
I'm just going to post a grounding question in the Lightspeed thread...
Thanks
Andrew
One meter only has two options, 750Vac and 200. The other two range from 250/200/20/2Vac
I did as you said. There is a reading for a split second and it quickly drops away to nothing.
Hi,
now that you know there is only undetectable AC on the output, switch to DC voltage and again starting at the highest voltage scale measure the output offset. If both of these show very low voltages then connect your source and measure again.
The increase is due to the source or due to the method of interconnecting.
At this stage I cannot tell if the method of interconnecting is the result of a grounding fault inside your power amplifier.
Some one else may be able to give you a series of tests to carry out that can identify the location of the problem.
now that you know there is only undetectable AC on the output, switch to DC voltage and again starting at the highest voltage scale measure the output offset. If both of these show very low voltages then connect your source and measure again.
The increase is due to the source or due to the method of interconnecting.
At this stage I cannot tell if the method of interconnecting is the result of a grounding fault inside your power amplifier.
Some one else may be able to give you a series of tests to carry out that can identify the location of the problem.
Less feedback makes the sound more relaxed to me, which
you may translate as open. I like it that way myself.
Since you are reducing the resistance on the Sources of the
Jfets, it's probably wise to stick with 6.8 mA Idss, which will
keep the temperature down on them.
Thanks Nelson
But as I told: the Mini F5 does not drive the tweeters as I wanted, because there must be a roll of in the higher frequencies.
I am waiting for delivery of the 2SK2013/2SJ313 to try.
I have finished moving my workshop and can (re-)start activities soon.
Franz
Hi Andrew
When I completed the amp I checked the offset over a few days, and at different operating temperatures. It never really went above 10mV without a source. I'll check again with a source connected.
I have a ESP ground loop breaker I could use in the Lightspeed if you think this would help. The F5s power supply ground is connected to the safety earth terminal with a CL60 in series
Here how the 1kHz square wave looks like (on my old Heathkit OS2) with the actual, unsatisfiying version of my Mini F5
Too restricted bandwith.
Franz
An externally hosted image should be here but it was not working when we last tested it.
Too restricted bandwith.
Franz
These readings were taken when the amp is cold
No source connected
Right channel
Peaks at 60mV and drops down to 27mV over a couple of minutes
Left channel
drops down to 14.2mV
Lightspeed and CDP connected, volume set at lowest setting
The readings are slightly lower than before, but the offset was slowly decreasing on the amp, so I guess there is no change.
Interestingly...when I rotate the volume control the right channels offset decreases from 27 > 8mV and the left increases from 12 > 37mV
this indicates no DC blocking capacitor on the F5 inputs.
The variable source impedance/resistance is affecting the DC quiescent currents flowing in the input stage and results in a variation in the output offset.
The Lightspeed has a high output impedance. Is the F5 happy with a high source impedance?
Maybe I should open a separate thread for the Mini F5 trials?
As I have my scope now available I just tried different bias settings between 0.4 and 1.2A at +-15V.
No big change, rolloff in the highs regardeless of the bias settings. It seems not to be a matter of current, much more of the limited Vds.
Now, I definitely wait for delivery of the 2SK2013/2SJ313.
Franz
As I have my scope now available I just tried different bias settings between 0.4 and 1.2A at +-15V.
No big change, rolloff in the highs regardeless of the bias settings. It seems not to be a matter of current, much more of the limited Vds.
Now, I definitely wait for delivery of the 2SK2013/2SJ313.
Franz
use your scope to look at the HF roll-off.
use a 1kHz sinewave that will result in ~2 to 3Vac at the output of the amplifier, with a 100r resistive load.
Set up one channel with the input signal at near full height of the screen.
Set up the other channel to monitor the output. Set it up to match the height of the input trace.
Now start increasing the sinewave test frequency.
It does not matter if the voltage of the sinewave varies a bit as you change frequency.
Now look at the difference between the output waveform and the input waveform.
If the frequency response is exactly -0dB and the phase is also exactly 0degrees then the output will lie exactly on top of the input trace.
If the frequency response drops by 0.5dB the output signal will be noticeably different from the input signal. A 6% drop in relative trace heights is -0.5dB. You can see that on the screen.
If you get to 100kHz and still can't see the output signal dropping relative to the input signal then that tells you the bandwidth is wider than 100kHz.
If the difference between the signals is ~12% (output = 89% of input), then that is -1dB
If the output appears to fall to 71% of the input height then you are at the F-3db frequency.
use a 1kHz sinewave that will result in ~2 to 3Vac at the output of the amplifier, with a 100r resistive load.
Set up one channel with the input signal at near full height of the screen.
Set up the other channel to monitor the output. Set it up to match the height of the input trace.
Now start increasing the sinewave test frequency.
It does not matter if the voltage of the sinewave varies a bit as you change frequency.
Now look at the difference between the output waveform and the input waveform.
If the frequency response is exactly -0dB and the phase is also exactly 0degrees then the output will lie exactly on top of the input trace.
If the frequency response drops by 0.5dB the output signal will be noticeably different from the input signal. A 6% drop in relative trace heights is -0.5dB. You can see that on the screen.
If you get to 100kHz and still can't see the output signal dropping relative to the input signal then that tells you the bandwidth is wider than 100kHz.
If the difference between the signals is ~12% (output = 89% of input), then that is -1dB
If the output appears to fall to 71% of the input height then you are at the F-3db frequency.
No, Andrew
Thank you for the hints, but:
When the 1kHz square wave is looking like this I showed, I dont have to take more care for the measurement.
My ears are telling me the same as the square is showing.
I will use the 2SK2013/2SJ313, first only one pair per output.
And then listen and measure again.
Franz
Thank you for the hints, but:
When the 1kHz square wave is looking like this I showed, I dont have to take more care for the measurement.
My ears are telling me the same as the square is showing.
I will use the 2SK2013/2SJ313, first only one pair per output.
And then listen and measure again.
Franz
Last edited:
Hi Tripmaster,
I have a CDP-lightspeed-F5 setup as well. I had the same problem initially (quiet when shorted, hum with source connected). Adding a shield to the interconnect eliminated/minimized the problem (shield connected to amp side, open on source side, seemed to work slightly better). Also, if the LSA is not powered there will be some hum/noise that is not there when it's powered.
As for the DC offset, mine also varies with the LSA. Perhaps it has something to do with the shunt resistor being in parallel with R10? I've experimented with removing R10 (LSA must be connected and powered to maintain ground reference) and noticed that DC offset varies even more.
Hi Alazira
That's interesting. I'll make up a pair of interconnects tonight and see if it helps. I have the parts to build a DCB1 so if there is an impedance mismatch this should helpfully improve the situation. I attached a crocodile clip to the star ground point on the LSA last night and probed various ground points to see if there was a grounding problem, but it didn't make any difference. I'll do the same with the F5 tonight. The hum is very quiet, and not that much of an issue, but its there all the same...grrr
Regards
I have a similar problem with my 4 channel Mini F5:
As soon as I connect the inputs to the active crossover (P09 Rod Elliott)there is some buzz. Also when the crossover is disconnected from mains.
I already tried different grounding topologies within the crossover and the F5.
As the Mini F5 is in double mono, I also tried with both tweeters on one side and distributed left and right psu.
Not solved up to now...
Franz
I also use the Lightspeed and I have floated the Grnd point (via the outer shield connection of the phono plugs) with a 100R above the chassis earth, as per usual - the buffer has the same 100R "lifter" and the amp uses one of those CL-60 thermistors for the same purpose - very quiet indeed - once had a problem with a USB Dac but one of those digital isolation tranformers solved that.
Perhaps this may be of assistance ....
Perhaps this may be of assistance ....
I just isolated but not solved my hiss problem:
One of the four channels seems to have a grounding issue or some other problem.
As soon as I connect the input signal wires of one of the other three amps to the input of the "problem" channel, buzzing is starting.
Maybe an undesired contact to the chassis somewhere?
Franz
/Edit
I am using input caps in front of the Mini F5. The problem arises only on the signal cables, not on the gnd connections.
One of the four channels seems to have a grounding issue or some other problem.
As soon as I connect the input signal wires of one of the other three amps to the input of the "problem" channel, buzzing is starting.
Maybe an undesired contact to the chassis somewhere?
Franz
/Edit
I am using input caps in front of the Mini F5. The problem arises only on the signal cables, not on the gnd connections.