So I upgraded the Nord Rev D buffer board in my Hypex NC 500 based amp with Sparkos opamps and 15v Sparkos voltage regulators. I decided to measure the DC offset at the binding posts with no load and nothing connected to the inputs. My meter started at 20 mV and over the next 10-15 minutes settled out around 350-400 mV. The range would fluctuate by some 30 mV consistency after 15 minutes and never stabilize. Obviously this seems like a very high number and it concerns me. I never checked the amp’s DC offset before the upgrade.
Should the amp be under load for this measurement? I can pick up an appropriate resistor if needed. Also, should I have the amp’s inputs connected? Thanks for any help guys. I don’t want to hook up my speakers until I know the amp is safe. The voltages check out fine and the regulators are holding a steady 15vdc +/- on the buffer board.
Should the amp be under load for this measurement? I can pick up an appropriate resistor if needed. Also, should I have the amp’s inputs connected? Thanks for any help guys. I don’t want to hook up my speakers until I know the amp is safe. The voltages check out fine and the regulators are holding a steady 15vdc +/- on the buffer board.
Last edited:
Hi,
If you don't have a oscilloscope to see what's going on at the output, yes, put a resistor as a load (4, 8ohm or something around this), short the inputs (or connect them to the pre-amp with volume to zero) and measure again.
Class D amps normally have some very high frequency spurious unfiltered voltage at the output, but the average voltage, which is what the DC multimeter measures, should be around 0mV (+/-50mV is ok). Unless this high frequency, if present, is interfering with the multimeter measurement - a oscilloscope would be better for this analysis.
Anyways, 400mV is too much indeed, so you need to resolve this issue somehow.
If you don't have a oscilloscope to see what's going on at the output, yes, put a resistor as a load (4, 8ohm or something around this), short the inputs (or connect them to the pre-amp with volume to zero) and measure again.
Class D amps normally have some very high frequency spurious unfiltered voltage at the output, but the average voltage, which is what the DC multimeter measures, should be around 0mV (+/-50mV is ok). Unless this high frequency, if present, is interfering with the multimeter measurement - a oscilloscope would be better for this analysis.
Anyways, 400mV is too much indeed, so you need to resolve this issue somehow.
It would be very interesting to see the noise and distortion measurements before and after these up grades. If you make the measurements, please post them here.
I have some resistors coming in tomorrow, and I will put the dummy load on the amp and short out the inputs and measure DC offset once again. I am hoping for a different measured result. All I have for equipment is a multimeter so any further testing, if the problem persists, will be problematic for me. As I understand the amplifier, it is direct DC coupled to the output so there are no coupling capacitors on the output. The capacitors on the buffer board are there I believe for filtering high frequency noise.
You should short the inputs for this measurement.
There is always a little input current for the input stage, and if you don't terminate the input it can easily drift to some undefined value.
If you want you can also terminte the output, but it is not necessary to use a low resistance dummy load as this is without signal. Just any old low ohm resistor will be enough for this test, 100 ohms for instance will be fine and does not need to be high power.
Jan
There is always a little input current for the input stage, and if you don't terminate the input it can easily drift to some undefined value.
If you want you can also terminte the output, but it is not necessary to use a low resistance dummy load as this is without signal. Just any old low ohm resistor will be enough for this test, 100 ohms for instance will be fine and does not need to be high power.
Jan
Designers spend a lot of time and effort carefully selecting the best op amps for use in their designs. Arbitrarily rolling op amps without very carefully considering all of the specified and unspecified characteristics of the op amps used seems unlikely to yield performance gains. Looking at the data sheet for an op amp you will find pages of specs including input offset currents and voltages. Adding to that the change to the physical layout, by adding a circuit board, will have some effect on the performance. I recently came across this video on the subject. Unless you are measuring before and after, as a designer would do, you won't get an accurate assessment of the modification. It is very easy to "upgrade" one parameter while down grading five others without knowing they were important.
The manufacturer specs this particular opamp as a plug in replacement for the exact board I have in my unit-his board. Just an FYI. You can use the Sparkos I have or Acoustic Imagery as they are two optional upgrades for the Rev D board. I did not add a circuit board. Not sure you understand my post or what I did at this point?
Last edited:
I will be taking my amplifier to a local designer and we will measure it together if the DC offset is truly high. I plan to measure the DC offset today with the inputs shorted and a load on the outputs.
Last edited:
I was not aware that Hypex spec'd this alternate op amp.
My point was that the Sparkos op amp is an added PCBA that is physically quite different from the original tiny op amp package. This adds trace loop areas and capacitive and inductive coupling paths to other components that did not exist in the original design.
I get it. It's all for fun. Have fun with it. My point is that these are complex devices with non-obvious circuit sensitivities. For instance changing the geometry of the traces can modify the EMI emissions and susceptibility. These board layouts are fine tuned due to the fast edges, high voltages and magnetic fields present. If you have the equipment it will certainly be interesting to see the before and after test results. I have a pair of Hypex modules sitting around, so I'm curious.
Understand your point now. All true. My amp was made by Nord Acoustics. They use the Hypex and other Class D OEM modules and then add their own buffer board with upgrade options. You certainly know more than I about design considerations and such.
Nord wanted me to buy new buffer boards, just like the ones I already have, but fully loaded with the same exact parts I purchased. The cost of the Nord fully loaded boards are more than 2X the cost of me doing the work.
Nord wanted me to buy new buffer boards, just like the ones I already have, but fully loaded with the same exact parts I purchased. The cost of the Nord fully loaded boards are more than 2X the cost of me doing the work.
Last edited:
Upgrades can get tricky. I helped sort out my friends Twisted Pear Audio DAC when he installed upgrade boards. Sometimes board versions and instructions get out of sync and things don't work. If it was easy it wouldn't be a hobby!
OK with load resistors on the outputs and shorting the inputs the DC offset only measured 4 to 6 mV. Looks to be all OK. Seems in order to get an accurate measure for DC offset with these class D amplifiers you do need to short the inputs and put a load on the outputs.
Big thank you to all for your help!
Big thank you to all for your help!
Ok, when I measure the DC offset with my preamp hooked to the amp’s inputs and speakers connected the DC keeps rising just the same way it did with the input open, not shorted. Just keeps rising slowly in the exact same manner. With the inputs shorted the DC offset is rock solid stable at 6 mV. Readings all taken from my Fluke 115 multimeter.
I am thinking my meter is picking up other “noise” and not giving me accurate DC measurements when the inputs are not shorted. What say you folks? I never measured the DC offset in this Class D amp before my upgrade. This may be normal with my multimeter as only the shorted input numbers are accurate. Not sure.
I am thinking my meter is picking up other “noise” and not giving me accurate DC measurements when the inputs are not shorted. What say you folks? I never measured the DC offset in this Class D amp before my upgrade. This may be normal with my multimeter as only the shorted input numbers are accurate. Not sure.
Last edited:
Hi! Maybe there is some DC coming from your pre amp and there is no DC blocking (or cap is not good) in the input of your class D amp.
Try to check pre-amp output (must be absolute zero) and amp cap input. If preamp output is not absolute zero, it means there is not output DC block or the cap is not working fine. If this is the case, or you fix it or you need a good DC block at the power amp input.
Also inspect the PCB, maybe when you desoldered/resoldered, there was some spurious conductive leftover between tracks or isles.
Try to check pre-amp output (must be absolute zero) and amp cap input. If preamp output is not absolute zero, it means there is not output DC block or the cap is not working fine. If this is the case, or you fix it or you need a good DC block at the power amp input.
Also inspect the PCB, maybe when you desoldered/resoldered, there was some spurious conductive leftover between tracks or isles.
Preamp has output coupling caps and no DC is measured on the outputs of the preamp. Good point on the rechecking of the boards. They look clean and good.
My multimeter on the amp’s output with preamp on the inputs. Inputs not shorted. DC runs down slowly from 40 mV to 0 then starts to run up slowly! Seems like a multimeter issue measuring unshorted inputs.
Did you try to measure the DC at the power amp input with the pre-amp connected?
Just to see if the power amp is not "producing" this DC at its input (this could come from a leak in the PCB or from a bad capacitor).
Just to see if the power amp is not "producing" this DC at its input (this could come from a leak in the PCB or from a bad capacitor).
Does not appear to be an input capacitor on the buffer board. I did hook up the amp to the speakers and shorted the inputs and the DC offset measured 4-6mV stable. It seems the inputs must be shorted to get a stable result.