Thank you for explaining even further . . .
Sorry for confusing everything. 1, .5, .25, .125, etc. is not linear of course. I knew spl to power and therefore perhaps volts was a log relationship. But I assumed from the beginning that 1, .5, .25, .125 was linear in my brain
for some reason and the .707 did not fit.
Got I now. The next post I make should be my test results and the f3 and slope, hopefully.
Sorry for confusing everything. 1, .5, .25, .125, etc. is not linear of course. I knew spl to power and therefore perhaps volts was a log relationship. But I assumed from the beginning that 1, .5, .25, .125 was linear in my brain
for some reason and the .707 did not fit.Got I now. The next post I make should be my test results and the f3 and slope, hopefully.
Volts max was at 29 Hz - 18.40, .707 of that (13.0) was at about 37 and at 12.84 volts. The rest of my results are:
Hz-volts AC
10-.367
20-6.79
30-18.24
40-10.30
50-5.17
60-3.19
70-2.59
80-2.46
90-2.38
100-2.25
125-1.74
150-1.20
175-.775
200-.488
I used some cheap (not metal oxide) 1 or 2 % tolerance resistors to provide resistance for the amp. I don't know if this was correct
. I thought with no resistor(s) the amp would get fired. They were two 10 ohm wired in parellel to give me a 5 ohm resistance.
What does this information mean
Hz-volts AC
10-.367
20-6.79
30-18.24
40-10.30
50-5.17
60-3.19
70-2.59
80-2.46
90-2.38
100-2.25
125-1.74
150-1.20
175-.775
200-.488
I used some cheap (not metal oxide) 1 or 2 % tolerance resistors to provide resistance for the amp. I don't know if this was correct
. I thought with no resistor(s) the amp would get fired. They were two 10 ohm wired in parellel to give me a 5 ohm resistance. What does this information mean
A couple of points about measuring
I believe you are measuring voltage. Therefore dB calculations must employ the 20Log formula. This leads to steps of double voltage equaling 6dB. The power will equal V*V/R.
If the calculations were done with the resultant power figures then 10Log is correct, and double power happens at +3bB.
There seems to be a big difference between the requency response of AC meters, though my experience of using them to measure audio frequencies has only applied to digital meters.
Some, Eg my current Fluke, have an excellent frequency response, certainly as good as any AC voltmeter in audio measurement kit. Some others are a disaster.
It is not difficult to test this. It can be done by measuring your audio source directly, and making a chart. If there is only a small error, perhaps of a dB or so, then a correction figure (coefficient for linear, or sum for dB) can be included in the calculation from the overall measured figures.
Another, simpler to many, method is to avoid using the meter's own rectifier.
A simple rectifier can be made using a small signal diode and capacitor, Eg 1N4148 and 0.1uF. Put the diode in series with one probe, and the cap across the meter. It doesn't matter which way, but the reading may indicate a negative voltage.
The voltage read will be the "peak". To calculate the RMS value, multiply by 0.707. This rectifier configuration will not work at small signal levels, so make sure you've got a volt or so.
I believe you are measuring voltage. Therefore dB calculations must employ the 20Log formula. This leads to steps of double voltage equaling 6dB. The power will equal V*V/R.
If the calculations were done with the resultant power figures then 10Log is correct, and double power happens at +3bB.
There seems to be a big difference between the requency response of AC meters, though my experience of using them to measure audio frequencies has only applied to digital meters.
Some, Eg my current Fluke, have an excellent frequency response, certainly as good as any AC voltmeter in audio measurement kit. Some others are a disaster.
It is not difficult to test this. It can be done by measuring your audio source directly, and making a chart. If there is only a small error, perhaps of a dB or so, then a correction figure (coefficient for linear, or sum for dB) can be included in the calculation from the overall measured figures.
Another, simpler to many, method is to avoid using the meter's own rectifier.
A simple rectifier can be made using a small signal diode and capacitor, Eg 1N4148 and 0.1uF. Put the diode in series with one probe, and the cap across the meter. It doesn't matter which way, but the reading may indicate a negative voltage.
The voltage read will be the "peak". To calculate the RMS value, multiply by 0.707. This rectifier configuration will not work at small signal levels, so make sure you've got a volt or so.
Jimmy:
I think this is the setup John, (Dhaen) was referring to. Of course, you can put the diode at the end of the probe, and hook it up to the amp.
All this stuff is available for really cheap at any Radio Shack for a couple of bucks. Since this is AC, I don't believe it makes too much difference which end of the diode hooks up to which-but you can switch it if you want.
Reason I am going through this is that huge peak at 30 Hz-wow! Just want to double check. Is there a switch on that amp for bass boost or anything like that?
Radio Shack part numbers, (from old catalog)
Diode 276-1122 ($1.19)
Capacitor 272-1069 ($0.79)
As you can see, this setup is quick and easy to do.
You might want to extend the setup out to 400 Hz. If the voltages fall below 1 volt, then go from 20 Hz up to where the voltage goes below 1 volt, then reset the volume up ward and measure the voltages out to 400 Hz in reference to 100 Hz.
For instance, you are reading now 2.25V at 100 Hz. If that holds, then when you get done with all the lower frequency measurement, boost the voltage so it reads 10 V at 100 Hz. Then go up to 400 Hz. Sort of break the test up into two parts.
Like I said that hump is quite substantial and I want to make sure that it is genuine.
As it stands now, taking 2.5 Volts as "normal", (located at 60 to 90 Hz), for your amp, the 30 Hz peak is 17dB above normal!
I think this is the setup John, (Dhaen) was referring to. Of course, you can put the diode at the end of the probe, and hook it up to the amp.
All this stuff is available for really cheap at any Radio Shack for a couple of bucks. Since this is AC, I don't believe it makes too much difference which end of the diode hooks up to which-but you can switch it if you want.
Reason I am going through this is that huge peak at 30 Hz-wow! Just want to double check. Is there a switch on that amp for bass boost or anything like that?
Radio Shack part numbers, (from old catalog)
Diode 276-1122 ($1.19)
Capacitor 272-1069 ($0.79)
As you can see, this setup is quick and easy to do.
You might want to extend the setup out to 400 Hz. If the voltages fall below 1 volt, then go from 20 Hz up to where the voltage goes below 1 volt, then reset the volume up ward and measure the voltages out to 400 Hz in reference to 100 Hz.
For instance, you are reading now 2.25V at 100 Hz. If that holds, then when you get done with all the lower frequency measurement, boost the voltage so it reads 10 V at 100 Hz. Then go up to 400 Hz. Sort of break the test up into two parts.
Like I said that hump is quite substantial and I want to make sure that it is genuine.
As it stands now, taking 2.5 Volts as "normal", (located at 60 to 90 Hz), for your amp, the 30 Hz peak is 17dB above normal!
Attachments
sreten said:Use of a linear frequency scale confuses matters : suggest using -
10, 12.5, 16, 20, 25, 32, 40, 50, 64, 80, 100, 125, 160, 200 Hz etc.
Hey, why not? These are one third octave measurments. If you want to extend it out to 400 Hz, add on 250 Hz and 320 Hz and of course, 400 Hz.
Or you can go linear, just as you did, and we can do the math later. You have more linear measurements, so they might be a little more exact. The main thing is to check the accuracy of the measurements, make sure that 30 Hz peak is genuine, and get a reading out to 400 Hz. Then we can tell what your crossover slope is for sure.
By the way, going by the measurements as they are now-and they very well might be accurate-your crossover slope is as follows.
200 Hz is 0.4 of an octave above 150 Hz. You go down from 1.2V @ 150 Hz to 0.488V @ 200 Hz. That is 7.8 dB. If we drop 7.8 dB in 0.416 of an octave, then we drop 18.7 dB in a full octave. Note: use the inverse key on the calculator.
Given the measurements and approximations of our measuring equipment, this would seem to translate into an 18 dB per octave slope. However, let's recheck our measurements and extend the frequency out to 400 Hz just to make sure.
200 Hz is 0.4 of an octave above 150 Hz. You go down from 1.2V @ 150 Hz to 0.488V @ 200 Hz. That is 7.8 dB. If we drop 7.8 dB in 0.416 of an octave, then we drop 18.7 dB in a full octave. Note: use the inverse key on the calculator.
Given the measurements and approximations of our measuring equipment, this would seem to translate into an 18 dB per octave slope. However, let's recheck our measurements and extend the frequency out to 400 Hz just to make sure.
dhaen said:Yes KW, that's the idea I was trying to explain
The 0.1uF value assumes that the meter has a high input impedance 1M -10M ohms, (as all digital meters I've come across do). If you find a fall-off at the LF end, just increase the cap value.
Thanks, John.
Hmmm, possible falloff ar low end. Okay, then pick up Rasio Shack part # 272-1069 capacitors, wire them in parallel, and run a 10 Hz or 20 Hz tone. If there is any change from the readings there with one capacitor, then you have a rolloff at the bottom end. If there is no change, then everything is fine.
By the way, just about every digital multimeter I have come across also, no matter how inexpensive, has an input impedance greater than 1 meg, so this is just added precautions.
Just note the circuit as shown by KC will under-read by
0.6 V and will not register anything below this voltage.
Why you don't trust a digital multimeter AC measurements
has me quite perplexed, my basic DM has a 200mV range
setting and you'd expect it to be accurate at 50Hz.
I'd suggest checking the signal source with the meter,
to look for any variation, if there isn't any leave it at that.
Measurements at 1/6 of an octave would be best, these are ;
10, 11.2, 12.6, 14.1, 15.8, 17.8, : 20 .... etc.
sreten.
0.6 V and will not register anything below this voltage.
Why you don't trust a digital multimeter AC measurements
has me quite perplexed, my basic DM has a 200mV range
setting and you'd expect it to be accurate at 50Hz.
I'd suggest checking the signal source with the meter,
to look for any variation, if there isn't any leave it at that.
Measurements at 1/6 of an octave would be best, these are ;
10, 11.2, 12.6, 14.1, 15.8, 17.8, : 20 .... etc.
sreten.sreten said:
So if he uses the illustrated setup and his meter reads 2.0 V, it really is 2.6V? That is good to know.
Maybe it would be good just to record the readings straight off the meter, then we can add the extra voltage when figuring the slope. That is, if he decides to use the setup at all.
He might just decide to compare the AC meter readings with the illustrated setup's readings at 30 Hz and 400 Hz, correct for the .6V, and if they check, that is that.
Digital AC voltmeters can start going off at 1,000 Hz or so I am told, 400 Hz is getting near that.
sreten said:
So would I. But that enormous peak of 17 dB at 30 Hz makes me want to check all possiblities. I have to admit, I didn't expect that.
Thank's Kelticwizard and John.
I will get a Fluke voltimeter and write down my results at the frequencies you (Kelticwizard) suggested. I figure it would be wise to try the Fluke before using the diode and capacitor setup, no?
I used this Fluke to measure inductance of my inductors at mostly 300 Hz and once I remember measuring an inductor at 100 Hz and it matched up with the inductance at 300 Hz. And then later I used an inductance meter to measure the inductance and it matched up to mostly +-1% or less.
I will get a Fluke voltimeter and write down my results at the frequencies you (Kelticwizard) suggested. I figure it would be wise to try the Fluke before using the diode and capacitor setup, no?
I used this Fluke to measure inductance of my inductors at mostly 300 Hz and once I remember measuring an inductor at 100 Hz and it matched up with the inductance at 300 Hz. And then later I used an inductance meter to measure the inductance and it matched up to mostly +-1% or less
.Jimmy:
Oh, so you are not paying for the Fluke? YOu already have it?
Because I wouldn't go out and pay the money just for this test.
I would point out that the parts to do the illustrated setup are only $2 at Radio Shack. Also, Radio Shack has a cheap analog multimeter which will be all you need for $12, if you have to buy new stuff. I really do think that the meter you have now is probably accurate. Just that the immense peak at 30 Hz is blowing me away.
You are going to laugh at this, but do you have an alternate amp to check this on?
Did you say this sub amp was from a Carver Sunfire sub? If it is, that might explain a lot. there might well be a rational explanation for that 17 dB peak at 30 Hz.
Also, is there a switch to make this amp flat response?
Oh, so you are not paying for the Fluke? YOu already have it?
Because I wouldn't go out and pay the money just for this test.
I would point out that the parts to do the illustrated setup are only $2 at Radio Shack. Also, Radio Shack has a cheap analog multimeter which will be all you need for $12, if you have to buy new stuff. I really do think that the meter you have now is probably accurate. Just that the immense peak at 30 Hz is blowing me away.
You are going to laugh at this, but do you have an alternate amp to check this on?
Did you say this sub amp was from a Carver Sunfire sub? If it is, that might explain a lot. there might well be a rational explanation for that 17 dB peak at 30 Hz.
Also, is there a switch to make this amp flat response?
I don't own a voltimeter. The voltimeters; one is my friend's and the $150 Fluke model is some one I know. I'm going to do the test when I'm done writing this, with the fluke.
Good thing you mentioned the alternate amp thing. I have actually been doing something very stupid and perhaps bad or unnecessary at the very least. I have been going threw my computer audio output to my HK amp via RCA's and then from that HK amp to my sub amp. My sub amp has RCA imputs.
The amp, is an amp that goes in a Acoustic Research 315HO (AR1 new model) speaker. It says on the back: "Sunifire Technology . . . manafactured under licence from the omnipotent Bob Carver and the almight Sunfire Corp."
No switch, that I can see.
Good thing you mentioned the alternate amp thing. I have actually been doing something very stupid and perhaps bad or unnecessary at the very least. I have been going threw my computer audio output to my HK amp via RCA's and then from that HK amp to my sub amp. My sub amp has RCA imputs
.The amp, is an amp that goes in a Acoustic Research 315HO (AR1 new model) speaker. It says on the back: "Sunifire Technology . . . manafactured under licence from the omnipotent Bob Carver and the almight Sunfire Corp."
No switch, that I can see.
My reads with the Fluke voltimeter are:
10 .059
12.5 .152
16 .429
20 1.091
25 2.43
32 2.300
40 1.740
50 .897
64 .520
80 .476
100 .464
125 .389
160 .249
200 .132
250 .058
320 .018
400 .005
For some reason at 25 Hz the volt kept going up then down over and over again from about 2.44 to 2.20 volts AC. Volts DC are all over the place at this frequency also, but remain at .142 volts DC for all other frequencies measured. I don't know if that means anything, probably not I'm thinking.
10 .059
12.5 .152
16 .429
20 1.091
25 2.43
32 2.300
40 1.740
50 .897
64 .520
80 .476
100 .464
125 .389
160 .249
200 .132
250 .058
320 .018
400 .005
For some reason at 25 Hz the volt kept going up then down over and over again from about 2.44 to 2.20 volts AC. Volts DC are all over the place at this frequency also, but remain at .142 volts DC for all other frequencies measured. I don't know if that means anything, probably not I'm thinking.
sreten said:
I'd suggest checking the signal source with the meter,
to look for any variation, if there isn't any leave it at that.
A) I agree. Unhook one of the RCA plugs from the CD player to the the amp, touch the leads to RCA plug, and run a few tones through to check. The tones should all be of approx equal strength except maybe for some under 20 Hz. Otherwise, something is amiss.
Pay special attention to the 25 Hz, 32 Hz, 20 Hz, 60 Hz, and 100 Hz range. Let's make sure they are all the same reading coming off the CD player.
B) The next test I would ask you to do is: simply disconnect the resistors, attach meter leads to one channel, choose any tone or combo of tones, and turn the amp all the way up and measure the voltage. In other words, what is the voltage at full output?
C) The next thing I would ask you to do is find out the amperage of the unit. Find out what the fuse current rating is, if it helps. Look at the electrical info at the back of the amp. Anything.
If that big 15 db or so peak remains, you might have a sub amp that is adaptable to a real world sub cabinet, as long as the loudspeaker is constructed by someone who knows what he is doing.
In other words, you might have a sub amp matched to a certain speaker. However, you can home build a speaker that it is talored to the amp. If you do, matching the subwoofer and the amp togheter will give you really improved performance over store bought subwoofers purchased to work with any amp.
The sound output on my motherboard audio out is more or less linear.
20 - 17.45 - .424
25 - ??? - .434
32 - 19.68 - .443
40 - 19.64 - .447
50 - 19.26 - .450
64 - 18.44 - .452
80 - 17.92 - .454
100 - 17.06 - .454
125 - 17.29 - .455
160 - 14.30 - .455
200 - 11.14 - .455
250 - 7.68 - .455
320 - 3.26 - .455
400 - 1.055 - .454
The first reading is my the amp at the gain set to the maximum setup (computer volume was at 40%, don't know if that matters).
The second reading is only the computer output through one RCA channel. I don't know what's going on at 25 Hz, I can't get a reading. And at 100 Hz the reading seems to dip. Actually from 100 and lower I don't think the readings are good. At 110 Hz the reading is 18.45 volts. The readings at 100 and below seem to stay at a high voltage and level off and then a few seconds later drop off to a lower voltage which is recorded above. As if the amp was being pushed to it's limit.
At less than max output there's still a peak, but I think I have a similar sub and I'm going to put it in a similar sized sealed enclosure as the amp was designed for. Or maybe I will put in a aperiodic vented box instead of a sealed box.
Got any idea about the x-over slope
20 - 17.45 - .424
25 - ??? - .434
32 - 19.68 - .443
40 - 19.64 - .447
50 - 19.26 - .450
64 - 18.44 - .452
80 - 17.92 - .454
100 - 17.06 - .454
125 - 17.29 - .455
160 - 14.30 - .455
200 - 11.14 - .455
250 - 7.68 - .455
320 - 3.26 - .455
400 - 1.055 - .454
The first reading is my the amp at the gain set to the maximum setup (computer volume was at 40%, don't know if that matters).
The second reading is only the computer output through one RCA channel. I don't know what's going on at 25 Hz, I can't get a reading. And at 100 Hz the reading seems to dip. Actually from 100 and lower I don't think the readings are good. At 110 Hz the reading is 18.45 volts. The readings at 100 and below seem to stay at a high voltage and level off and then a few seconds later drop off to a lower voltage which is recorded above. As if the amp was being pushed to it's limit.
At less than max output there's still a peak, but I think I have a similar sub and I'm going to put it in a similar sized sealed enclosure as the amp was designed for. Or maybe I will put in a aperiodic vented box instead of a sealed box.
Got any idea about the x-over slope
Jimmy154 said:
The readings suggest a sixth order alignment with a peak around 29Hz.
Very fine measurements should be taken around 30Hz to find the peak.
And then rapid crossover to the mid units, certainly heading to 36db/octave.
The peak is still at least 14dB.
What does concern me, but its impossible to tell, is if the
amplifier is also manipulating output impedance. In fact its
not impossible to tell, you have to change the load impedance,
e.g. double it and see what happens. If the results are different
then impedance is being manipulated and it gets very complicated.
Its possible the peak is caused by innapropriate load impedance
and can be varied by changing the load of the amplifier.
What is still not clear is if the subwoofer amplifier contains any
passive components for the mid units crossover. I must stress
that with high order slopes as seems to be the case here you
cannot approximate, you have to be exact.
sreten.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.