Hi John
Add on to my last reply.
Reference R40 and C17 filtering effect in the speaker amplifier (low pass).
Just did a quick calc. The roll off (3db) point is about 2.2kHz which will be definitely NOT HiFi.
If I get a chance to-morrow I will try to do a simulation (I have a simulator on another computer) on these values and see exactly what it looks like.
Cheers Bob
1 Like
Hello Bob,
Re : Also there is a signal level at the collector of Q3 of 3.7V with a reference to note 4 (which I am not privy to) which is probably close to what is across the primary. 3.7 / 2 = 1.85:1 which is in the ball park.
The 3.7V you are reading is in brackets, which means that is an RMS value.
The 13.0V that is not in brackets is the DC voltage
The ā4ā is terminal 4 on the transformer ā not note 4
Note 4 has nothing to do with this and reads
Re HiFi ā this application is definitely not. It is for voices and alarm tones played over horn speakers.
Re : I had a quick look and could not see anything that looks like those resistance figures (ignore the tertiary winding, it would not be used), they are all up in the 100s of Ī©
While you were away, I went and bought a transformer for testing. I believe the figures you were looking at (if you were looking at the Altronics web site) are impedances, because I bought a 10k : 2k transformer and measuring the dc resistances I get 513 ohms : 58 ohms
This is a tiny transformer that only measures 15mm x 15mm x 10mm ā it does not list a frequency response.
https://www.altronics.com.au/p/m0212b-10k-ohm-2k-ohm-audio-coupling-transformer/
I think these types are low power ā in the 450mW range ā hence the small size
What power do I need ?
If I go by the size of the transformers in the comms stations, they are more like the M1000 size below
I think I have a 600 : 600 , So I will measure that and see what I get.
I can the go and try both of them and see what we get.
There is a much bigger transformer
M0709 ā¢ 600 Ohm / 600 Ohm Bridging Input Transformer
25mm x 29mm x 27mm
Frequency response 20Hz to 30 KHz
https://www.altronics.com.au/p/m0709-600-ohm-600-ohm-bridging-input-transformer/#/
Or even bigger
M1000 ā¢ 600 Ohm / 600 Ohm Telephone Isolation Transformer
41 x 30 x 33mm
300Hz to 10 KHz
https://www.altronics.com.au/p/m1000-600-ohm-600-ohm-telephone-isolation-transformer/
Re : You need to feed the transformer via a capacitor to get rid of the DC component ā¦.
Is this inserting the electrolytic cap in the +ve line between the amplifier and the transformer ? And the other leg of the transformer to ground ? Can you provide a quick sketch of the circuit please ?
Re : This OP amp may not have enough power handling capability for this job ā¦.
I am copying the op amp circuit and building it with larger components, mainly to be able to provide a larger, hand operated pot, instead of the tiny one on the breakout and I am using an LM358. I understand I can use up to 32V with the LM358. And yes, 12V would suit me better too.
Is using an LM358 going to give me enough power using a 12V supply ?
Re : At what point to you intend sourcing your audio on the MP3 device, at the headphone socket / terminal block or the speaker output.
I intend to use the audio jack and just one channel
2 Likes
Hi John
Got to simulate the low pass network comprising R40 (2k2) and C17 (33nF).Not without drama. I use a Mac by preference but my simulator is on a Windows laptop which I recently gave a new lease of life with a SSD and Windows 10. I then had to re-register my sim on a ānewā computer which entailed a search for my original order number. Anyway I succeeded and here is the result:
As you can see not HiFi but as you say the whole thing is voice frequency anyway so doesnāt have to be.
I had realised that as I saw it mentioned somewhere
Sorry, missed that but obvious when pointed out.
Yes I was and I know I was looking at the impedance on this site. Altronics donāt publish DC Ī© figures but these figures I referred to were from Element 14 who do publish them. You were able to actually measure them as you have one.
We should make one thing clear here. You may be aware of this or may not be.
The 10k to 2k transformer you apparently have will only look like 10k and 2k on the following conditions. The 10k will only look like 10k when the 2k side is terminated with 2kĪ© and the 2k will only kook like 2k when the 10k side is terminated in 10kĪ©. The actual impedance seen on the primary side will be the value of the secondary termination multiplied by the impedance ratio which is the square of the turns ratio. In your case I have assumed the line termination to be ā 33Ī©
Thus looking into the 2k side we see 33/5 (5=imp ratio)= 6.6Ī©, Looking into the 10k side we see 33 * 5 = 165Ī©.
Now 6.6Ī© is far too low for the LM358 output and 165 should be just about bearable so to try this connect th e 10k side to the amp output via the 4.7ĀµF electrolytic with +ve side toward the opamp.
The next thing to consider is the level required. The transformer has a turns ratio of 2.24:1 (ā5) which means for afinal output of 1.5V we need 1.5 * 2.24 = 3.36VRMSat the op amp output. This is a peak to peak of 9.5VP-P which is right on the op amps range with a 12V VCC (1.25V to 10.75V). 10.75V and 165Ī© relates to 65mA which is over the OP amps capability and will clip the audio when the max is reached.
The 1:1 transformer will be no good to you with the LM358 as the load will then be 33Ī©, too low.
You really need an amplifier which is capable of driving a speaker. If you are looking to build that amp on another board you maight look at the LM380. Connecting the transformer with the 2k side toward the OP amp (2k trans primary) will present a load of 6.6Ī© which is within speaker range. You will also get a 2.24 times voltage step up so your 1.5V should be achievable easily.
You will have to increase the electrolytic capacitor to 10F or more. Maybe 47 or even 63ĀµF
Not quite sure what you mean here. The OPamp output will be at VCC/2 (6V with 12V VCC) so you need to isolate this. Connect the capacitor between the OP amp OUTPUT pin (or board connector marked āOUTā) and transformer, the other end of the winding to ground.
Refer above. Even increasing the supply voltage will help the P-P capability but will not affect the clipping when max current is reached.
You can combine both channels with 5k to 10k resistors as done on the MP3 trigger board. Might sound a bit better.
There is one thing you could try fairly quickly.
I am assuming you are using the spark fun trigger DEV16892.
Use the whole MP3 trigger board including the speaker amp. Connect the 2k side of the transformer, I donāt think you need capacitors here as it is designed to connect a speaker directly. You can try it without connecting it to the line but donāt forger to terminate the 10k side with a 33Ī© resistor.
I believe the volume can be set with software, I am not familiar with that.
It is possible this may be all you need. If you need a physical pot that is possible but you will have to hack the board. I have never sighted this physically so donāy know how difficult this will be.
Edit: Sketch
Cheers Bob
1 Like
Wow !
I very much appreciate your expertise Bob and ability to explain the details in terms I can understand.
I am also learning from you as we go through this 
-
List item Thanks for confirming with the simulation.
-
List item Transformers - 10K to 2K DC measurements are 520R to 61R
600R to 600R Isolation transformer DC measurements are 37R to 32R
Re - Connect the capacitor between the OP amp OUTPUT pin (or board connector marked āOUTā) and transformer, the other end of the winding to ground.
Yes this is what I have done. Sorry for my poor question.
Using the 10K to 2K transformer worked, but it was still lacking some oomph. I could adjust the volume pot from zero to full, without clipping or distortion. But I think it may have been getting close to clipping or distorting - not sure.
I tried the 600R to 600R transformer with the 4.7 uF electrolytic in the +ve line to the transformer. As you said, this provides way too much and is clipping or distorting. The volume is also high when at the almost minimum position.
Can we put a resistor (say 150R) in one of the transformer lines, before the 33 ohm resistor termination ?
OR
You seem to have solved the MP3 Trigger output problem by putting a cap in the line to the transformer. What will putting a cap in the positive output line of the transformer do ?
Remember - the nominal DC resistance that I measure across the Page line is 33 ohms in series with a 100 ohm pot (they call this line balancing) and when we have one of the old tone generators across the line, the DC resistance then drops to about 8 or 12 ohms. Then, depending on how far away from the terminating resistor you are, you will also measure a bit more cable resistance, but that does not usually add much resistance - maybe 10, 20 or 50 ohms at max.
We are getting close to the solution, and I would prefer to not alter existing boards. So is there a way to use the 1 to 1 transformer by adding a component or two somewhere ?
Regards
John
1 Like
Which āJohnā are we talking to here. Have we 1 person with 2 names or 2 different people.
Would be handy to know.
Cheers Bob
1 Like
Hmmm ā¦
Looks like one person with two names - donāt know how that happened.
Do you know if there is a way I can fix it ?
āJohnā
1 Like
You have probably registered twice in the past and somehow you replied with the āotherā name. Core will probably have a solution.
will digest the last reply tomorrow.
Chers Bob
1 Like
Now we have established we are talking to the same āJohnā we can continue.
The DC resistance of the transformers does not mean a lot. It can be handy to identify windings but it is the turns ratio that is the important bit. The DC Ī© ratio may not be relevant as different size wire could be used for different windings but will give you some idea, that is all.
The AC impedance (also expressed in Ī©) presented to a circuit is more important than the DCĪ©. For instance that transformer in the speaker amp has a quoted DC R of 1.5kĪ©. I estimate the AC impedance presented to the line in that circuit would be something like 150kĪ©. Just a bit different and like I have been trying to explain depends on the turns ratio (NOT the DC Ī© ratio) and the terminating resistance in the secondary.
Once again you are getting worried about DC resistance,
When you connect the tone generator you are measuring the 10Ī© secondary of the transformer in parallel with the line DC resistance so I would expect about 8Ī©.
Putting these 2 quotes together makes me suspect you are doing a listening test. DO NOT connect your device to the line AND have a 33Ī© terminating resistor connected. You will be double terminating. The 33Ī© termination is for āstand aloneā testing and ideally done with the aid of an oscilloscope (to view any clipping). Failing that you could use the AUX input of another audio amp which should have an input imp about 50K. In this case use the 33Ī© resistor to simulate your line in a stand alone test.
That is because the trimpot used will almost certainly be a linear curve. For audio volume use the curve should be logarithmic or as sometimes called āaudio logā to more suit the human ear.
You can try this with the 600Ī© transformer. Donāt use a single resistor, keep it balanced by using say 100Ī© in each side (total 200Ī©). This will help the OP amp by increasing the primary Z from 33Ī© to 230Ī©. This should lower the level on the line by something like 60%.
I am a bit confused here. Since when have I suggested that in regard to the MP3 trigger. You still have not confirmed which trigger unit you are using. I am just assuming it is the spark fun device. Core sell several. The only mention of using the complete (including audio class D speaker amp) as described in the last paragraph of my previous reply. Please read as I donāt feel like continually repeating. All other trials have been assuming the LM358 amp. Please do not mix up as this is difficult enough remotely.
Nothing. There is no positive line as the line is balanced. The only way this would be required is if any DC conditions would be changed by the transformer resistance. As everything seems to be transformer I donāt see that as a problem so donāt bother.
See above.
On the subject of measuring across a balanced line. If using an oscilloscope in a 'stand alone" test using the 33Ī© termination the ānormalā methods are OK.
If measuring across the balanced line however REMOVE the mains earth from the instrument and make sure the case is not touching any other earthed instruments. This makes sure you donāt get any funny phantom earth connections likely to cause irregularities or other confusion.
Personally I hate earthed oscilloscopes as this can introduce things that are not really there. I have mine powered via a short 300mm length of mains cable with the earth wire visibly hanging out of the socket so others can clearly see there is no mains earth connection on the case.
Cheers Bob
1 Like
Yes I am doing a listening test.
I am not sure we are on the same track about the 33 ohm resistor. Just to confirm -
Yes, in my stand alone testing, I have a 33 ohm resistor across the line.
In a live situation the 33 to 133 ohm resistance across the line will always be there. This is the way the system is designed to operate. I need to add the MP3 Trigger with the Op amp and the transformer to the system and everything needs to remain in place and connected.
Is this the way you are seeing this ?
Re -
I am a bit confused here. Since when have I suggested that in regard to the MP3 trigger. You still have not confirmed which trigger unit you are using. I am just assuming it is the spark fun device. Core sell several. The only mention of using the complete (including audio class D speaker amp) as described in the last paragraph of my previous reply. Please read as I donāt feel like continually repeating. All other trials have been assuming the LM358 amp. Please do not mix up as this is difficult enough remotely.
I am sorry if I ask the same questions - I donāt mean to.
I have attached a sketch of what I have set up on the bench. I am concerned about introducing a speaker amp signal onto the Page line, because that line is usually, only 1.5Vrms and the 33R resistors across the lines are only 1/4 watt resistors. Each Remote Comms Station has its own speaker amplifier to power the speakers connected to it.
As we have too much gong onto the line with the setup as shown in the sketch, I feel it is better to reduce this rather than add a speaker Amp to the line that could blow the 1/4 watt resistors.
I will go and add the resistors as per the sketch.
In the meantime, could you confirm the questions in the sketch please ?
John
1 Like
Hi John (we have still got the new John)
Yes. maybe just the way I interpreted your message but I was concerned that you may have the line AND the 33Ī© terminating resistor connected but it seems you donāt.
The sketch shows the resistors going in the correct places with only the line connected. Al good.
I considered without these resistors the load presented to the LM358 was a bit too low. The addition of these resistors will increase this load to 233 - 333Ī© which is a bit better. This will also reduce the voltage on the line by about 60+% which you may or may not have up your sleeve. Reducing this value would change the division ratio but also reduce the LM258 load.
My concern re clipping is twofold. 1) Clipping due to excessive voltage swing P-P. This can be overcome by increasing the VCC. and 2) Clipping due to reaching maximum current the LM358 can supply. The only solution here is to use another amp.
Delete all previous reference to using the speaker amp on the MP3 switch. I have been assuming the wrong MP3 device. It is not until now that I have had any confirmation as to which device you are using and even then it has been by comparing pics. This reinforces my earlier statement re not enough initial information can lead to confusion and unnecessary research. It would have been a bit of help to provide that info some days ago. Like I said, like pulling teeth.
If you find the first stage of the LM258 breakout board having too much gain you can reduce it using the method I outlined before. I am not going to repeat it.
If you are using the pot as a āvolumeā control you would be better off removing the trimpot and extending the connections to a log or audio log pot. The control will not be so cramped up at the bottom end as you have found.
Cheers Bob
2 Likes
Hi John
Hold it.
Stop everything.
Damn it.
Just looked up the data sheet on that LM358.
It is LMV358 and would you believe the max VCC is 5.5V so I think there is no way I believe that IC is going to get you 1.5VRMS onto that 33Ī© line.
Sorry about that. I have been using 258/358 amps for years with a VCC of 30+V and did not see why this 358 should be any different except for surface mount.
Try it at 5V but I donāt think it will make it. I hope I am wrong but you are going to have to do the experimenting. Start by not fitting the 100Ī© resistors.
I think I should retire after that mistake.
Cheers Bob
2 Likes
Thanks Bob,
DONāT RETIRE !! I need as much of your help as I can get !
I have only ever talked about the MP3 Trigger, because that is all I have.
And I asked, but later confirmed that the LM358 can handle 12V as VCC. As you said up to 32 V I think.
Hence the expanded Op amp has the LM358 and is being powered by 12 V and this is also the maximum VCC that the MP3 Trigger takes.
Now I am confused ā¦
When you say āI was concerned that you may have the line AND the 33Ī© terminating resistor connected but it seems you donāt.ā
When you say āLineā are you referring to the line out of the transformer and going to the Comms Station ? In this case, the āLineā has the 33 ohm resistor across it. This cannot be changed, it is how the system works.
I have added 82R resistors in the two lines out of the transformer (because that is all I had and this keeps it balanced, as you said).
I am gong to change the 82R resistors for 100R resistors and also change R1 in the Op amp to 51K
As a side note, when I measured the resistance on the line out of the transformer and connected to the Comms station, with the 33R resistor across the line and with no resistors in each leg line, the reading on the multimeter was jumping around. When I had the 82R resistors in the legs, the reading was steady. Hence, in my belief, one more reason to have the 100R resistors in the legs out of the transformer.
I shall go and test with different resistances and let you know what I find.
Regards
John
2 Likes
Hi John (we are back to the original āJohnā
Core have several MP3 triggers in stock. They all have part numbers. Two are spark fun I guessed the cheaper one. I was wrong. The part number would have helped.
That is correct. But the chip in that board is LM V 358, or that is the data sheet linked in the attached docs for that item. The āVā is the difference
We are having a bit of terminology confusion. Yes the ālineā I refer to is the line you have going around all the stations and yes I am aware it has the 33Ī© resistor / 100Ī© pot across it at all times. The āotherā 33Ī© resistor I refer to is the terminator to simulate this ālineā when testing stand alone.
This will be ok to prove concept. The difference is not great. 100Ī© came out of mid air as a typical value.
This will be a good move as a total gain of 100 is a lot especially as you get enough from the headphone jack to drive headphones. If you find the first stage clipping or overloading you could terminate the headphone output with 33Ī© to simulate headphones.
I see no reason for that unless you had some audio there or something.
Recheck that data sheet with the V in the type number.
Also check the board to see exactly what chip has been fitted.
Cheers Bob.
PS I might not retire yet. Got to try to keep the old grey matter going for a while.
1 Like
Not quite correct. this device has the ability to provide either line level signal or drive headphones, but not both. the device is supplied as standard as line level output and you need to alter jumpers on the board if you want to change it to headphone output.
God to hear you will be around for a good while yet ! 
Regards
John
1 Like
Hi John
I assume from this you are using āline levelā. A delightfully vague term. Vast difference between āprofessional line levelā and ādomestic line levelā. I will also assume ādomesticā which should be about 900mV P-P or, if a sine wave about 320mV RMS. There are a lot of āassumptionsā here and it would be nice if you could measure it but to do this successfully you will need an oscilloscope. You cannot measure music or other complex waveforms with a DMM, or an analog meter for that matter.
Just looked at the codec specs. It appears the output is about 1V P-P so looks like domestic line level. That just about answers this question.
Have you checked the data sheet and had a look to see which 358 (LM or LMV) is actually fitted to that board yet?
Cheers Bob
1 Like
Hello Bob,
Thanks for that info. It is helpful. A long time ago when I started this, I connected the output of the MP3 Trigger directly to the Comms station with the33 R across the line.
1 Like
Doubt if that would have worked. Not enough level and the output of the MP3 thing is unbalanced, one side is ground. The good thing about having a balanced line in the first place is you can run it over long distances with minimum interference (telephone lines for instance).
Cheers Bob
PS. My other question re actual IC???
1 Like
Re - PS. My other question re actual IC???
I presume you are asking which one I am using - right ?
I am using the LM358P with 12Vdc power
1 Like
John
I an getting fed up with trying to identify boards from pics. EXACTLY what is that OP amp breakout board you are using. It does not look anything like the one Core have stocked as BOB09816 that you originally quoted. That one is supposed to have LMV358 which is surface mount. The one you just quoted as LM358P is through hole.
If this is indeed the case and you have a different board I am going to feel very sad indeed after looking at the wrong description and circuit all this time.
As I said before accurate description and as much info right from the start might just get results. I think I have mentioned pulling teeth before.
Cheers Bob
1 Like
Bob, I am not sure that all my answers are getting through.
Sometimes I answer from the email and sometimes from the web site forum.
So letās try this again from the web site -
Re
John20131
Re - PS. My other question re actual IC???
I presume you are asking which one I am using - right ?
I am using the LM358P with 12Vdc power
As mentioned earlier, I am using the LM358P.
I donāt know what other number I can give you, other than to say as I said before, I have taken the circuit of the Op amp breakout SKU: BOB-09816 and made my own PCB with larger components so I can have a hand operated volume control. As part of converting to larger components I used the LM358P. Yes this is through hole because my PCB is through hole. And as mentioned earlier it is powered by 12 Vdc. see below.
I hope this gives you the information you are after.
Regards
John
1 Like
