- Posts
Hello all,
I thought that the owners of the SSTRAN AMT3000 would be interested in how I measured the final input power in my unit. This is not that hard to do, so if you are curious, try it. All you need is a voltmeter and a calculator. You will need a schematic for this to make sense and I will not provide one since it is copyrighted.
. . . . . . . . . . . .Hello all,
I thought that the owners of the SSTRAN AMT3000 would be interested in how I measured the final input power in my unit. This is not that hard to do, so if you are curious, try it. All you need is a voltmeter and a calculator. You will need a schematic for this to make sense and I will not provide one since it is copyrighted.
. . . . . . . . . . . .Description of measurement of the SSTRAN AMT3000 AM transmitter final input power.
For those who just want the results, I measured the final input power to be 91 milliwatts which was not affected by the load. I measured the same power with three different antennas and a 50 ohm dummy load.
The 100 mW. legal limit applies to the DC input power to the final RF stage. It has been common practice to measure transmitter input power as the DC input power. So for this unit, we need to measure the voltage across the final and the current through the final. Power is then P = VI.
My description is keyed to the AMT3000 schematic rev. 05-1. The SSTRAN schematic is copyright SSTRAN, so if you need one, contact them. I will not provide it.
There may be some interpretation problems about the final stage (see below). I used Q5 as the final stage and did not include the base bias current. We can indirectly measure the current through Q5 by measuring the voltage across R8 (VR8) and, using ohm’s law, calculate the current through Q2.. This current is the sum of the currents through Q1, Q4, and Q5. The Q1 current is obtained by measuring the voltage across R3 (VR3) and applying ohm’s law. This, subtracted from the current through Q2 gives the current through Q4 and Q5. Since Q4 and Q5 are symmetrically biased, IQ4 = IQ5 and the current through Q5 is one half the total current.
Summary:
IQ5 = 1/2 ( VR8/R8 – VR3/R3 ) = 7 milliamps in my unit.
The voltage measured at the emitter of Q5 ( VQ5E ) to ground = 2.03 volts.
The voltage across Q5 (C to E) = +15 volts – VQ5E.
The power is P = ( +15 – VQ5E) x IQ5 = 12.97 volts x 7 milliamps = 91 milliwatts.
If one wants to count Q2 as being part of the final stage, P = 15 volts x 7 milliamps = 105 milliwatts (excluding the bias current from Q1). I exclude Q2 from being part of the final stage because it does not process the RF signal. It serves as a constant current sink which is controlled by the audio. Common practice is to measure transmitter final input power under conditions of no modulation, therefore Q2 power does not add to the RF signal. Q2 provides a constant current bias for Q4 and Q5 which is included in the calculation for final stage power. The only power Q5 has available to produce output power is the voltage across Q5 x the current through it and it is reasonable to use this figure. The power used by Q2 does not produce RF output and can be excluded. (This is similar to excluding the filament power in a tube since the filament power, though necessary for the tube to function, does not contribute to the output power)..
For those who want to see how a similar circuit works, search MC1496 balanced modulator/demodulator data sheet and application notes.
Neil
Total posts : 45366This post from Radio8Z is most interesting, and I have a question that goes to the heart of the AMT3000 DC power input:
Given your finding that your transmitter measures at 91mW, what can be done to bring the DC input upward an additional 9mW?
Total posts : 45366From the schematic it appears that if you make R9 smaller the bias current to the final will increase increasing the power. I have not tried this so this is just my best guess.
Also, things such as modulation linearity and headroom may change so proceed with caution.
Neil
Total posts : 45366Yeah but will raising that bias throw off the conductive curve enough to dampen performance rather than improve it.
Depends on if the active device in that circuit has the characteristics and curve at that increased voltage point to continue to amplify while not adding junction noise or some other problem like saturation.
Be neat to find out though.
RFB
Total posts : 45366Radio8Z, your comment seems to be absolutely correct.
I also had estimated that R9 (47k) was the key to bias voltage setting for Q3, the point at which audio enters the modulator, and the Manual Description of the Modulation Section states that the arrangement at this juncture is designed to maintain 100mW “as close as practicable.”
Since my older AMT3000 was already on the bench for the checkout blogged about yesterday, I decided to see what would happen if I made the bias somewhat adjustable.
R9 got removed and other parts, as necessary, un-soldered for desired circuit modification. In place of R9 I used 33K in series with a 20k trimmer pot, so I had a bias range of about 1.7 to 3.0V. But the only setting that gives the desired modulation is about 1.82V, as indicated in the schematic. Boosting the bias voltage very quickly degenerates the audio until it becomes very distorted. That probably indicates the “loss of linearity” you referred to.
The designer was very accurate in determining the exact parts in the transmitter.
Total posts : 45366RFB wrote “Yeah but will raising that bias throw off the conductive curve enough to dampen performance rather than improve it.”
I am not familiar with the term “conductive curve”. Do you mean the load line or the device characteristic curve?
Carl,
Changing the bias can alter the saturation and cutoff of the final stage. Ideally the Q point is centered between these two points on the load line and changing the bias moves the Q point up or down along the load line. This is what I meant by “head room”.
In plain English, this means that the final needs to be able to swing as far positive as it does negative at 100% modulation lest there be distortion. I have found both by experiment and by simulation that the AMT-3000 is fairly sensitive to proper RF loading. Improper loading gives distorted audio and RF.
Neil
Total posts : 45366Thank you for the shared experience, radio8Z.
The test cited above was made feeding the dipole antenna which was by the workbench left-over from the Big Talker shortwave project, and I loaded the AMT3000 into it for the trial with the bias adjustment. The poor results are mentioned above.
But your comment about the RF load affecting the modulation sent me back to try again with the now famous dummy load. A totally different result!
I now have the transmitter bias adjustment wide open, putting 3.0V bias Q3 base rather than 1.82V, and the modulation is sounding good, although it needs to be driven harder to reach 100%.
But the most interesting thing of all is that adjusting C5 (RF output) very dramatically PEAKS the modulation!
S5 Output Filter set for 82uH
I’ll try the antenna again and tinker with C5
Total posts : 45366Three cheers for Radio8Z
What the heck, make it four cheers!
Test antenna attached, 3.0V bias set, tuning of C5 centers and peaks the modulation; de-tuning C5 causes modulation distortion.
I had originally used a DVM to set the RF output into the antenna, so I’ll have to compare and see if the Voltage peak achieved by tuning the RF unmodulated, is the same setting as when peaking the modulation with C5.
To remind, the purpose of this modification is to allow fine tuning the 100mW input to the final RF stage.
Now the question that reveals my engineering limitations…..
What bias setting would be right on the money at 100mW?And, for idle curiosity, how many milliAmps are being generated with a bias of 3V ?
Total posts : 45366Sounds to me like this thing has inadequate coupling isolation between the final and output load. It should not be affecting that final to the extent of attaching a load and throwing off the conduction curve to cause dampened modulation.
Perhaps it should be operated as it was designed to be operated from the onset, and find another unit capable of loading up to a proper load without adverse effects of everything before that load.
Sounds like your having fun learning what’s really there vs what was touted. That’s a good thing.
RFB
Total posts : 45366RFB said – “Sounds like your having fun learning what’s really there vs what was touted.”
That statement contains one part truth and two parts mistaken.
Yes to having fun. This academic exercise came about when I read radio8Z’s finding was 9mW short of limit, and I wondered if it would be possible to achieve a fine adjustment.
What was touted is the same as what’s really there, a part 15 transmitter with a legal input to the final, designed to load small antennas and within FCC specs.
Total posts : 45366Carl,
I am not sure how to predict by how much you should change the bias current but a good approximation should be to increase the bias current by 9% and check the results. This change shouldn’t change the final output VCE since the emitter voltage is fixed by a resistive divider.
On the other matter, there is no complete matching nor filtering network included in the final of the AMT-3000 since it relies on part of the antenna series inductance to perform this function.
The other thing that strikes me is the apparent constant input power in the situation where the load is changed. This indicates to me that very little power is being transferred to the antenna system. My reasoning is that the input power goes two places, to heat, and to the antenna. If changing the antenna and hence the power to the antenna produces little or no change in the input power then it seems this means very little power is going to the antenna. There could be higher order effects in play but this circuit is so simple that I doubt this.
With the high efficiency transmitter I designed there is a very large change in input power with a change in the load (output power). This is because 85% or more of the input power is going to the output.
Neil
Total posts : 45366The easiest way to increase the power to 100 mW is to replace R7 (510 ohms) with a 620 ohm 5% resistor. This will yield about 101 mW, but getting exactly 100 mW is not possible given standard resistor values and component tolerances.
Keep in mind that increasing power input from 91 to 100 mW is a 9.9% increase in power, but only a 4.8% increase in field strength so the benefit will likely not be very noticeable.
Due to the modulator transistor configuration, the AMT3000 circuit provides a constant power input regardless of the load from the antenna. The modulator transistor is a constant-current driver to the RF transistors. The current is determined by the DC bias on the modulator transistor. With the constant supply voltage and constant current, the input power is constant.
A simplified way to view the action is that the constant input power is divided between dissipation in the RF transistors and dissipation in the antenna load. With no antenna connected, all of the input power is dissipated in the RF transistors (this is safe by the way). With a tuned antenna connection, more of the power is dissipated in the antenna and less in the RF transistors
I want to underscore that the standard AMT3000 was originally designed with small internal, switch-selected inductors with the intent for use as an in-home transmitter. The small inductors are very lossy, but more than adequate for in-home broadcasting. When the external base-loaded antenna is used, the internal inductors are bypassed and the big low-loss external coil is in the circuit. Range improves dramatically.
This difference has been a point of confusion over the years. For those who want maximum range from the AMT3000, the external base-loaded antenna is the solution. If you just want in-home range, the basic AMT3000 will deliver. The range estimates for the AMT3000 with the two configurations have always been available on the SSTRAN web site under the title “About Transmitting Range and Antennas”.
Phil
Total posts : 45366Phil,
Interesting information about the AMT-3000. Thanks for posting.
Neil
Total posts : 45366Special thanks also from me, PhilB, for the interesting information about your great transmitter.
Getting to know our transmitters is part of the part 15 cult (culture).
Total posts : 45366With the high efficiency transmitter I designed there is a very large change in input power with a change in the load (output power). This is because 85% or more of the input power is going to the output.
i’m still interested in acquiring those designs even more so now i have a FIM, SA, and L/C meter
- You must be logged in to reply to this topic.
