The AM Forum

I picked up a Flex 5000 over the weekend at the Marlboro flea.  The price was right, so it followed me home....

Anyway, I put the thing on the air, hooked to my normal audio system that I use with the class E rigs.  The Flex has incredible frequency response, and is quite clean.  However, there was one problem - no negative peak limiter.  The result of this is the "folding back" AKA wavelets problem when one exceeds 100% modulation.  Such rigs don't really overmodulate - they fold back.   The resultant distortion is particularly annoying in standard AM detectors (fine in a sync detector)....

The flex has a particularly good low frequency response (down to single digits) so a negative peak limiter is very practical to use here, and there won't be any waveform "tilt".  So, I built a negative peak limiter followed by a 6 pole filter with the 3dB corner at 5.7kHz, and it works just great !!!

I can hit 200% positive modulation with 90% negative modulation no problem.  The audio when looking at the mod monitor looks identical to the class E rigs, and it sounds extremely similar as well.

The N.P.L. circuit is quite simple, and I will publish it here shortly (probably Tuesday - it's getting very late here).

I've heard a lot of Flexes and modified SSB rigs that have balanced modulators on the air - and they can sound really great except for that "folding back" problem when one hits 100% negative modulation.  The negative peak limiter takes care of all of this, allowing lots of audio on the carrier without the resultant "overmodulation" distortion.  The low pass filter prevents "splatter".  As I said earlier, this is the identical N.P.L. and filter as is used in the class E rigs, so the audio quality is not diminished at all.

Boy you work fast! Just got it home and now this! Great!

I can't wait - is a PCB in the works?

Here is the schematic:  Simple to build dead bug construction over a copper clad board (for RF immunity).  Maybe a board if there is enough interest.

(http://www.classeradio.com/negative_peak_limiter.jpg)

Steve
I could be interested in a entire unit.
Carl
WA1KPD

Basically,  the input network the latest pwm board.....   Yeah?

I had broken pieces out of that for my pwm and processor.   Although I used my assymetry board,  I use an npl on baseband audio as a stop gap.

--Shane
KD6VXI

pwb yes

I would buy a completed QIX made unit with SLR connectors on both ends. I would also be interested in a board. Seems to me there would be a rather large demand among us hams and the CB market for base stations would eat them up. Maybe AM broadcast stations would like them as well unless they already have a very high price equivalent?

I think it would be a must have item for Flex owners..

Steve,

Well, it looks like you may now be in the market to build a linear amplifier, unless it's just a hi hi FB 100 watt effort used as an exciter for the big rigs. This now opens up the whirl to hi-fi SSB and even CW.

Your perfect linear Segway would be developing a digital linear that runs 90% efficiency. Years ago there was a lot of talk about them being mainstream ham radio by now, but no ham manufacturers have stepped up to the plate.  We all still suffer with class B blow torches.

Be sure to make use of their version of "Pure Signal" software. This will make any ratty linear perform like a class A dreamboat.

T

I'd be interested in PCB or kit

.

I already use a CRL PMC300A neg peak limiter into my Flex 3000 but I would certaily be interested in a PC board for this project.
Don W9BHI

Does the circuit need to be removed when operating SSB or FM?

Hey, that's a good feature that may be worth adding.  Never thought of that because, for me anyway, I never operate any other modes.  However, that is not true of most folks.  I guess a Negative Peak Limiter "Disable" feature would be worth adding.  Leave everything else in line so the levels don't change - just disable the negative peak limiting.   Or alternatively provide a bypass feature.

I'll take input on this - might as well get it right the first time  :)

Regards,  Steve

I would be interested in a PCB.

Just for clarity.

This is an inboard modification?
or an outboard audio mic or line level device??

                      _-_-bear

Strictly outboard.

FYI:
For the Flex Series:
Flex 5000 has single(unbalanced) and balanced mike inputs.
Flex 3000 has single(unbalanced) mike input.
Flex 1500 has single(unbalanced) mike input.
Flex 6300 has single(unbalanced) mike input.
Flex 6500 and 6700 have single(unbalanced) and balanced mike inputs.

Just a thought; since many SDR-type rigs run off 12 volts DC, it would have been convenient to design your circuit to run off the same 12 volt line.

Count me in for a board

If you decide to move forward with this, it might be worth submitting it to QST for possible publication. That would certainly sell PCB's (and or kits) and would compliment the ARRL's recent interest in AM. For me , this would be the missing element in my audio chain.

I like the idea of the negative peak limiter but why would you need a low pass filter after it?
The audio bandwidth can be set in the Flex software.
It uses a brick wall low pass filter.

Don W9BHI

That's a fine question!  Not all rigs have a low pass filter (this circuit is designed to be used with other good rigs other than just SDRs), and also the filter shown in the schematic is "gentler" (6 poles) than the brick wall filter in the SDR and sounds a bit nicer to the ear.  In my own Flex, I actually set the transmit filter to 7.5kHz (each side of center), but the filter in the negative peak limiter, which is the same one I use in the class E rigs, has the corner set to 5.7kHz. 

At night, or if I need more filtering because there is a nearby QSO, I kick in a steep filter (an additional 8 poles, giving a total of 14 poles) which is part of the audio chain for all rigs in the shack.

The filtering in the SDR is a nice thing to have for sure (if you're got an SDR)  ;)  !

I have to say I am quite pleased with the performance of the receiver (it is absolutely amazing) and the transmitter.  I think it's the first commercial piece of ham equipment I've ever encountered which needed no modifications whatsoever and works perfectly in all respects right out of the box.

Steve,
I'm glad you approve of the Flex 5000A.
I have had mine for about a year now and couldn't do without it's great receiver.
Don't get me wrong, I still love the boat anchor stuff like my BC-1G and R-390A.
Some people just don't even want to hear about SDR radio's.
Too bad, I think it's the future of ham radio.(but what do I know)
I will be listening for you sometime on 3873 to hear how you sound.
Thanks for all of the work you do on this stuff.
Don W9BHI

 

I guess the Flex 5000A is at a point in its life cycle where owners are curious and going into the diddle/enhancement mode with their rigs. I've received several requests for the printed service manual over the last several weeks. Maybe we'll see some user mods over the next several months.

I think we will be seeing more Flex radios in general. While at the Orlando Hamcation I stopped by the Flex booth. They were demo'ing their new amp and I overheard Steve Hicks mention that they've got the adaptive pre-distorion working very nicely with the new amp. I suspect that it may be part of v2.0x

I also saw three Anan users put deposits on Flex radios.

Hi Steve,
  I am just guessing, but would that circuit output ideally be fed directly to the DBM in a rice box bypassing the audio section of the radio?

Bill

With many of the current transceivers on the market  today with DSP circuitry, they don't use a "balanced modulator". Like my IC-756 PRO II, I haven't used my mike input since I bought the rig. My audio input into the rig goes directly to the Accessory socket when there is a modulator input connection.

Steve, if you were to make a PCB, would you use surface mount components, or standard leaded components? One way or another I'm going to start gathering the parts. Also, I see there a several resistors that aren't standard values. Are their value that critical, or can you substitute standard values? I guess it would be easy enough to use some miniature trim pots to get them spot on.

Pete, my radios are older than yours.  ;D

Is sw301 a bypass for the filter? I would like to use this circuit with my ts440 that currently has a 7.5k filter. It would be nice to narrow things up when necessary.

SW-301 changes the output audio phase.

Thanks. After looking at the datasheet for the tl074 I think I see some light through the darkness.

Nothing surface mount (at least not yet).  It's just so much easier to build using through-hole components.

Steve,

you should talk to the clown that designed this thing....  WAY too many parts.

And those odd values!


:)

Sounded good on K2SDR today


--Shane
KD6VXI

Thanks for the report  ;) ;D  Good conditions on 75 this afternoon for sure!

Just so everyone knows what Shane is talking about - I was assembling this circuit today while I was in QSO, and "complained" about the designer (me) that there were too many parts in the circuit.  That's where a PC board will make it easier.  In reality, there aren't THAT many parts, but I was growing tired of soldering at that point.

Regards,  Steve

Thanks for the short chat last evening on 75. You mentioned that you have the "critical value parts" - let me know how much you want and I'll get some $$ off to you. That might be the only difficult part of this project.

Between the junk box and mouser I have everything but the1n34a. Im going to hit the local surplus store next week. I may end up trying some different substitutes and see if I can hear any difference.

how ironical is this ... the first hobbyest solid state electronic components were 1N34a germanium point contact diodes and a few similar transistors (2N109, CK722, 2N34, etc) .... so many projects were developed for these

nice work Steve and others

The negative clipper diode is not at all critical. 

Steve:  While I was listening to you, I had a friend in the shop.  Newly minted General class.

You keyed and started talking, and he picked his head up and looked at the screen.  When you made the statement about 'who designed the thing', he looked at me and said 'I want that!'.

Shortly after, he was on classeradio site taking in all he could...

Nice station, and it makes a great advertisement.  What where you running, class e or sdr?

--Shane
KD6VXI

In a pinch, sometimes you can use the base to collector leads of an old germanium transistor.  ;D

That was the class E rig.  The Flex is the receiver and will be the VFO for remote operation.  Otherwise, for local operation, I'll use my homebrew VFO.

Steve,

   I heard you guys on 3873 and then 3875 last night. Timmy was doing a job to the 3873 Texas "no traffic" net on 73. He made a good Old Buzzard transmission. :D

You were as loud as he, and comfortable copy at around 10-20 /9 on my Flex 3K.

I have a Flex 5K coming, also with 2nd RCVR and ATU. Should have it by this weekend...

After that I will have a Flex 3K and an Icom R-75 4-Sale.

Jim
Wd5JKO

For any of you who are going to build one, I created the shopping list. Those components noted with an * are critical values, and should not be changed according to Steve.

Value          Qty   Component
0.001   10   CAP
0.0015 *   1   CAP
0.0022 *   1   CAP
0.01           1      CAP
100uf   2   CAP
25uf          1       CAP
3300uf   2   CAP
47uf          1       CAP
1N34A   1   DIODE
1N4007   4   DIODE
LM7812   1   IC
LM7912   1   IC
TL072   1   IC
TL074   1   IC
2K POT   1   POT
5K POT   1   POT
560          2   RES
10K          3   RES
17.4K *   1   RES
2.2K          2   RES
2.87K *   1   RES
220K          2   RES
22K          2   RES
28K          1   RES
4.7K          2   RES
6.81K *   1   RES
8.45K *   1   RES
9.31K *   1   RES

Appreciate that , ordering today.

Are all resistors 1 watt? Maybe that is spelled out on the schematic but I writing this on the fly. In addition, would the first person who orders all the parts include an approximate cost?

Thanks Terry

It would also be great to see examples of your build. I have built all sorts of homebrew tube stuff but never solid state.

No, 1/4W is more than enough

Here's how I buy my standard components ...
http://www.ebay.com/itm/3120pcs-156-Value-1-4W-1-Metal-Film-Resistors-Assortment-set-1-ohm-10M-ohm/182366765338?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D41395%26meid%3D6f20dc258ad94235919a45e5f19652b3%26pid%3D100005%26rk%3D3%26rkt%3D6%26sd%3D182409997607

Good list!  I made some comments/quick edits.  I did not verify that every part in the circuit is there, but I did verify the critical parts and made notations where I edited.

Thanks!!

Value          Qty   Component
0.001   10   CAP (Edit: 2 are critical values - need to be very close to .001)
0.0015 *   1   CAP
0.0022 *   1   CAP
0.018 *     1      CAP  (Edit: this cap was missing from your original list)
0.01           1      CAP
100uf   2   CAP
25uf          1       CAP
3300uf   2   CAP
47uf          1       CAP
1N34A   1   DIODE
1N4007   4   DIODE
LM7812   1   IC
LM7912   1   IC
TL072   1   IC
TL074   1   IC
2K POT   1   POT
5K POT   1   POT
560          2   RES
10K          3   RES
17.4K *   1   RES
2.2K          2   RES
2.87K *   1   RES
220K          2   RES
22K          2   RES
28K   *      1   RES (Edit: this part is a critical value)
4.7K          2   RES
6.81K *   1   RES
8.45K *   1   RES
9.31K *   1   RES

Steve
I have several TL071's which I believe are 1/2 of the 072's. No reason I can't use 2 071's in place of the 072, right?

Peter - thanks for the info.  You always have the right info at the right time. Anybody have the circuit Peter is referring to?  Gotta be a simple  DC to DC converter - probably a switcher. It would be nice to be independent from the AC line in case one is operating under say a 12 volt battery.  That might be better put in a separate thread

Al

Add me to the list of ops with interest in purchasing a blank PCB.  

I had boards made for an old PE project from this guy last year.  I was very impressed with the quality.  His prices are very reasonable.  All he needs is a good clear photo print of the board layout.  

http://www.farcircuits.net/

Although I've made numerous projects using perfboard, it's a PIA.  An etched board makes a much nicer, neater project, and greatly reduces the possibility of a wiring error.   The possibility of a wiring error increases with complexity.  This circuit is complex enough where a wiring error is possible with perfboard.  That's why I'd gladly pay for a printed circuit board.

.

Great question ... My audio chain is balanced throughout, and I see that this is only "line" out. Not sure that it makes that much difference feeding the radio with unbalanced line out. Also have one more question - why two phase reversal switches, @ input and output?

Already started building mine.

1.  Add a phase splitter with a tl082 or 72 op amp on the output.

2.  Phase switch on in and out because when your audio signal goes through the Low Pass Filter you can have phase reversal.

--Shane
KD6VXI

I'll have to research that one - thanks.
Since I had several 074's, I decided to use those throughout. The unused sections could probably be used for the above. I don't think there would be any audible difference using "line" vs a balanced output. Thanks

You want the polarity (phase is an incorrect term) on the input so that the limiter is working on the side of the signal with the smallest amplitude. The switch on the output is required to ensure the side of the signal with the largest amplitude is modulating the transmitter in the positive direction. Without these two switches, much of the effectiveness of the processor could be lost.

Thanks for the education ...
Very much appreciated

DC coupling to the modulator, I suppose...
There is a need to check this point.

G.

Steve is correct.  The NPL shown works properly for the Flex and with class E rigs because they are set up such that positive going energy generates a positive peak.  This may not be the case with all transmitters, in which case a phase selector on the output would be required.  An easy solution would be to use a TL074 instead of a TL072 in the input, and use one of the other op amps in the TL074 as a unity gain inverting amplifier.   Then add a switch to select between the inverted and non-inverted output.

Sorry guys, I don't know what I was thinking.  The negative peak limiter ALREADY has an output phase switch included (I guess I anticipated this need, and put it into the design).  So ignore everything I said circuit-wise in the previous post  ::)

The circuit is all-ok as is.

Regards,  Steve

The latest release of PowerSdr for the Flex Legacy series now has the ability to flip the audio phase. The latest from Darrin Ke9ns is V2.8.40, and the phase flip feature is described below:

06/16/17 v2.8.0.40 Add Phase invert function (setup->Transmit->Phase invert) for your mic audio. To see the difference set the Display mode to Panascope, and toggle the Phase Invert while you transmit

This is good news!

Jim
Wd5JKO

This circuit could be made even more useful with an adjustable positive peak limiter for those who want to protect mod transformers and other components from excessive voltages.

Same perceived loudness and cleaner sound on most receivers.

Hi, Steve.  I don't remember if we've ever QSOd but I enjoy the PM-AM net on Sunday evenings on 3837. 

Thanks much for posting your limiter circuit.  I have a much modified FT-920.  I drive the balanced modulator directly from my own input xfmr-coupled op amp buffer and an external audio rack. 
A question was raised a few posts above that I was wondering as well: Your circuit is all DC coupled and in your full modulator you run all DC coupled.  Does your limiter need to be DC coupled to the balanced modulator to prevent flux balancing if input transformers and/or capacitors are used in the audio chain?  I can do that but I'd need to remove my present input circuit and rework it as DC coupled to the balanced modulator.  If DC coupling is not required, I can (happily) leave it alone.
Tnx es 73,
Chuck
NA3CW

With any asymmetrical limiting, DC coupling is definitely better.

There is a small DC shift that occurs with this type of circuit.  Any AC coupled elements will tend to try to "integrate" this DC shift over time.  If things like coupling capacitors, etc. are large relative to their terminating resistance, the time constant will be long and unless someone put a constant (negative peak clipped) signal into the system, in which case the negative peak clipped waveform will shift lower, all will work quite well.

The other thing that occurs with AC coupling if the audio rolloff is in the usable audio range is waveform "tilt".  This comes about because any low frequency roll-off will tend to act like a differentiator, and the leading edge of the clipped waveform will have a higher amplitude than the trailing edge.

That is why with plate modulated or other transformer coupled modulators, applying negative peak limiting after the modulation transformer/capacitor choke combination (if so wired) works much better than trying to put an asymmetrically clipped signal through the iron.

So, the bottom line is this:  If your transmitter has VERY good low frequency response (flat to lower than 10 or 15 Hz), waveform tilt and shift shouldn't be a problem with normal, varying audio waveforms.

Hope this helps  :)

All very true.

Broadcast audio processing pioneer Robert Orban (wonder if he's related to Viktor Orban) once wrote an article in which he suggested that, in order to accurately reproduce an audio processor's tightly-peak-limited waveform, an AM transmitter should have flat response down to 0.7 Hz -- and no, that's not a typo.

73,

Kevin, WB4AIO.

Yes, Steve, it does help. Thank you for the quick reply.  Your response wraps much better wording around what I imagined to be the situation.  The balanced modulator in the FT-920 is a run-of-the-mill IC balanced mixer.  The carrier level is determined by a DC imbalancing current.  At present the audio is AC coupled and rolls off at about 100 Hz so it's not an ideal signal path for asymmetrical waveforms.   

If I brought out a buffered DC-coupled signal pin to the outside world, I could create a full audio chain with frequency contouring, asymmetrical clipping and filtering, carrier control, etc, all in one package and have full control over the results.  I might even throw in a phase rotator.  :-)  Given that the FT-920 is not particularly amenable to experimentation, with all its surface mount components on the backside of the motherboard and a zillion hardwired connections around the edges, I may hold this project off until retirement.  My Yaesu is already quite non-standard with relay controlled filter bypassing, selectable audio sources, modified ALC, etc.   In the meantime I have some interesting design ideas to work out. 

Thanks again and 73,
Chuck
NA3CW

I ran into problems when I tried to AC couple a NPL into my otherwise DC coupled PWM.

My carrier would shift ALL. over the place!  Very unnerving!  Looked like controlled carrier gone wild!

Removed the DC carrier level control on the pwm and the associated capacitor on the audio  input and let my NPL and assymetry board (mine stretches positive peaks, the npl is a baseline prohibiter in my case) inject a DC level instead.

All is now happy in assymetry land.  I can stretch to 420 pct pos peaks and don't suffer from the carrier following the low frequencies.

Another way of explaining what Steve said!  ☺️

--Shane
KD6VXI

Two thoughts...

A board that fits into the D-104 base might be neat.

I know, or I think I know, that QIX doesn't mind using quad opamp packages.
But, if anyone is going to go and make a run of PCBs, my opinion is that duals
are far preferable since the quad DIP package is totally obsolete.

There are a whole lot of duals that are very nice for audio applications.
Even surface mount devices that can be put onto DIP adapters, if one wanted to
do so.

Phase flip might be implemented with a simple "see-saw" phase splitter, and a switch
to pick which side of the phase splitter you want as the output? That would be a single
transistor or JFET... cap couple of course.

In a past life I was Chief Engineer of KTWR, Guam.  While they have a different set of transmitters now, at that time we had four Harris SW100A's and one HCJB HC-100, all 100kw HF AM broadcast transmitters.  The SW100A's had 80kHz PDM tube modulators.   I designed and installed Dynamic Carrier Control boards for them.  Saved about $5000 per month in power bills.  When there was no audio they'd settle back to 50kw.  As audio arrived, they'd perk back up to 100kw to accommodate the modulation.  We didn't save money when music was playing but we did well when there was a speaker who had long, thoughtful pauses.  ;)

Since I had DC control of the modulators I could mix in carrier level, linearizing feedback, processed audio, and control scaling, into the same circuit.  As Shane related, that's about the only real way to pull off this asymmetrical processing, I think.  It should be straightforward to do a simpler version of this for my FT-920 balanced modulator.

Thanks to all for the great comments!

Chuck
NA3CW

Just about done building it, so now I have a question. Going to be feeding a Flex 6500. My PR-40 feeds a Behringer DEQ2496 which currently feeds the 6500 via a balanced audio connection. Not really doing much processing with the 2496, but it was the only thing I had that could serve as  Mic preamp. Since the limiter does not use balanced audio,  I'll use the "LINE" input on the 6500. I don't expect any difference in quality.

Am I correct in thinking that the limiter needs to be the last device before audio hits the 6500?

Here is my view.

An equalizer should only be used if there is some deficiency in the audio chain, or you're trying to generate an NRSC-type pre-emphasis curve.

Assuming you have plenty of mic gain, the limiter is usually all that is need for amateur radio AM and is placed just before the transmitter audio input.

I was not aware the Behringer DEQ2496 had a mic preamp, but by golly it does:

https://www.soundpro.com/catalog/documents/DEQ2496_brochure.pdf


Phil - AC0OB

So in my application the 2496 is a huge overkill, but it serves as a very nice preamp. I usually use the PR-40 which really does not need too much EQing. I have a few other mics that I use from time to time, which can benefit from some "enhancing", so the 2496 does that.

Once I get the limiter done I need to order a mod-monitor, something I've been meaning to do for some time. For me the recent Flex software update was free, so the $200 I saved can be applied towards the mod monitor. Thanks for the input.

In my case, I ended up using a Behringer multiband compressor, DSP9024, with its peak limiter disabled, feeding into the line input of my (older SDR-1000) Flex. I then use the leveler/peak limiter function of the Flex to do the final peak control. It works well, better than external peak limiters I have tried.

I have the leveler adjusted so it acts slowly (2ms attack, 5000 ms recovery, 500 ms hang) and never changes the gain more than 4 dB -- this just assures that I am in the "sweet spot" for the final limiter. I am not familiar with SmartSDR, but with PowerSDR 2.5.3, I have choice of "CPDR" (really a clean peak limiter misnamed as a compander) and "DX" (a more aggressive limiter that allows some clipping). Both work well at low settings. With DX I never exceed a setting of zero, and it still is as aggressive as a old-time Top 40 broadcast AM peak limiter. A CPDR setting of 2 is plenty.

In my opinion, the internal limiters work better than the Behringer's limiter function or the Aphex Dominator I was formerly using.

73,

Kevin.

Yes, the negative peak limiter (and its filter) should be the very last thing in the audio chain.

Look forward to hearing your report (and maybe you directly!).

Regards,

Steve