Larry East, W1HUE
Tucson, Arizona
(contact author)
Reprinted from the January 1997 ARCI QRP Quarterly
Before making any mods
described in this article, be sure to read the Notes and Corrections article.
The Index Laboratories’ QRP PLUS offers a lot of features in a small package. It has a continuous coverage receiver (1.6MHz to 30MHz) and a 5W (nominal) transmitter that covers all amateur bands from 160M through 10M. It is one of only a few commercially available QRP transceivers and has become very popular since its introduction. This is indeed a nice little rig, but not without a few faults – some technical and some more in the realm of personal preferences [1, 2]. The QRP PLUS is an evolving product as Index strives for continued product improvement. A "new improved" version of the rig is now being produced that has several major changes over the "original" model. Owners of the original model were offered the opportunity to have their rigs upgraded for about overhalf the price of the new model. I had one of the original models to which I had already made several modifications when the new model was introduced. After comparing my modified rig with a new model [3], I decided not to take advantage of the upgrade offer.
Many of the modifications that I will describe are applicable to both the new and original models, and some are model specific. From this point on, I will refer to the original model as the "QRP+" and the new model (including "upgraded" units) as the "QRP++". This is common practice among owners, but it is not an "official" Index Labs designation. "QRP PLUS" will refer to either model.
I have divided this article into two parts because of the large amount of material covered. This first installment will describe modifications that apply primarily to the receiver. The second installment will concentrate on transmitter modifications. Some of the modifications that I will describe have been adapted from the work of others and I will try to give proper credit in those cases.
Index made several "firmware" upgrades for the QRP+. The most recent version of which I am aware is designated "Rev 4D". The Rev 4C and 4D EPROMs contain several programming changes that affect the internal keyer and the switched capacitance audio filter (SCAF). How do you tell if you have this version or not? Simple: If you did not upgrade, you don’t have it! If you obtained your rig used and don’t know its history, a couple of simple tests will determine whether or not you have the latest EPROM: If the internal keyer speed changes in increments of 5 WPM, you don’t have it. If the keyer speed changes in increments of 1 WPM but an external keyer cannot be followed above about 27 WPM, then you still don’t have it (you probably have Rev 03).
NOTE added March, 2014: EPROM upgrades are apparently no longer available!
The small size of the rig can be rather intimidating the first time you remove the cover. However, if you have some building experience, a good soldering iron with a small tip and a little patience, you should be able to successfully perform the modifications that I will describe. A desoldering tool and/or good quality desoldering braid is also a necessity. I find that a desoldering tool is useful for removing the "surface solder" but that desoldering braid is a necessity when removing parts from double sided boards with plated-through holes. A magnifying lamp and/or a magnifying glass will also come in handy.
Two things that you should do right off the bat: Install some type of connector on the speaker cable, and install flat washers over the hold-down holes on the top PC board (the RF board). The primary reason for installing a speaker cable connector is to avoid having to replace the speaker when you eventually knock the cover off the bench and rip the wires out of the speaker! I drilled the rivets holding the small terminal strip to the speaker frame and mounted a small right-angle aluminum bracket in place of the terminal strip. I secured the bracket to the speaker frame with a sheet metal screw and epoxy, and mounted a mini phone jack in a hole in the bracket. The two wires from the speaker voice coil were then unsoldered from the terminal strip and soldered to the jack. I installed a mini phone plug on the cable connecting the speaker to the front panel phone jack. One of the speaker cable wires is grounded, so make sure that it connects to the grounded side of the jack if you go this route. An in-line jack and plug arrangement will also work; just make sure that they don’t hang down too far and touch something on the RF board!
The purpose of the flat washers is to prevent the lock washers under the board hold-down bolts from destroying the ground plane around the holes on the top board. The plating gets pretty chewed-up after several removal-reinstall cycles. I soldered zinc plated brass washers directly to the plating around the holes so that I don’t have to mess with putting them in place every time I remount the boards. I retained the lock washers under the mounting bolt heads – they now contact the flat washers soldered to the ground plane rather than the groundplane plating.
Here are a few things to keep in mind when working on your QRP PLUS:
The following modifications are be described in this installment:
This, of course, is not specific to the receiver, but it is an easy change to make and a good place to start. You should make this modification even if you make no others. You might be inclined to say: "Hay, I’m not stupid enough to connect a power supply up with the wrong polarity!" However, I know of at least one case in which damage occurred to a QRP+ when the owner attempted to inset the power connector with the power applied and accidentally touched the ground ring of the plug to the center pin of the power jack.
The rig does have reverse polarity protection of sorts; a reverse-biased diode is connected from the +12V power buss to ground (on the XMTR board). The idea is that a reversed polarity power connection will result in the diode conducting heavily enough to blow the fuse. There is a slight problem with this scheme, however – the diode (a 1N4005) is rated at 1A and the fuse is rated at 3A! There is a good likelihood that the diode will open before the fuse blows thereby removing any reverse polarity protection.
To provide "sure fire" reverse polarity protection, I replaced the wire between the power connector and fuse holder with a 1N5822 3A Schottky barrier diode. (Actually, I used an SK9935 because it was available locally.) Any leakage (it should be very small) through the series diode under reversed polarity conditions will be shunted to ground by the 1N4005. The small voltage drop across a Schottky diode – I measured 0.28V across the SK9935 while transmitting at 5W output – will not affect transmitter or receiver performance.
This is where my "QRP+ modification binge" began. Within the first three months that I owned the rig, the receiver input mixer had to be replaced twice (the first time, the IRF510 transmitter final was also destroyed). The second time I returned the rig for in-warranty repair, I received a phone call from Index saying that I must have a static discharge problem or possibly a problem with RF from a nearby transmitter. I decided I had better try to find out what was happening.
After posting a query on the Internet, I learned that several other QRP+ owners had experienced similar problems; one owner lost the mixer in the middle of a QSO during a wind storm. A look at the QRP+ schematic revealed that there is pretty much a straight shot from the antenna connector to the mixer input. Checking the schematics for my FT-301 and TenTec 509, I noticed that they had protection diodes at the receiver input; the QRP+ has no such input protection. Ah ha! Installing some protection diodes at the input of the mixer might solve the problem!
I installed four 1N4148 diodes in series-parallel at the output of the receiver input high-pass filter, as shown in Figure 1. I used four diodes rather than two in order to reduce any effect on the receiver IMD. However, the QRP+ IMD immunity is not all that great to start with, so using two diodes (in reverse-polarity parallel) would probably have no adverse effect.
Caution: Do not use PIN or defused junction rectifier diodes; use small signal Si switching diodes only!
Figure 1. Mixer protection diodes.
I originally put the diodes before the HP filter [4 ] but later decided that the series impedance of the filter’s capacitors would be helpful in limiting the current through the diodes. The diodes are mounted on the bottom of the RF board, between one solder pad for C49 (the end connecting to the relay) and the ground pad for C46. The diodes will certainty not protect the mixer from a very large discharge induced by a nearby lighting strike, but they should provide protection from moderate static discharges. To be completely sure, you should disconnect the QRP+ from the antenna when it is not in use!
Early production runs of the QRP++ had the mixer protected by an MOV device installed at the same location that I used for the protection diodes in my QRP+. In a fairly recent model that I tested, the MOV was missing. I have been told that Index decided that the mixer that they are now using in the QRP++ is more robust than an MOV and are no longer installing MOVs. However, if I owned a QRP++, I believe that I would remove the MOV, if present, and install protection diodes in its place. The mounting holes for the MOV (labeled "VR1" on the board) can be used for the diodes rather than mounting them on the bottom of the board.
I later realized that a static charge build-up on an antenna can easily reach a high enough voltage to destroy the protection diodes and then the mixer. After the protection and input mixer diodes are destroyed, another discharge can arc across the antenna relay and destroy the IRF510 MOSFET. This is probably what happened to my rig the first time I lost the mixer. Adding a resistor in the 10K to 20K range directly across the antenna connector allows any static charge to bleed-off quickly without doing any damage. I now place a 10K 1/2W resistor across the antenna connectors off all of my rigs and antenna tuners and have not experienced any problems from static discharges.
I noted several receiver "spurs" in my QRP+ on almost every amateur band, particular 80 meters. Many were too weak to register on the S-meter, but could be distinctly heard with the antenna disconnected. The loudest spurs – registering S-1 or greater – were at 3570.8, 3845.6, 7142.2, 14286.5 and 21161.4 kHz. A spur near 10.0 MHz was strong enough to make WWV difficult to copy at times. These spurs appear to be present in all QRP+ receivers, but the intensities vary somewhat from rig to rig (perhaps due to LO cable positioning – see below). Very few spurs were noted on the QRP++ that I tested [3], and the ones that were present were very weak.
Note: The following applies to a non-upgraded QRP+ only!
The intensities of these spurs can be greatly reduced (some eliminated entirely) by four simple changes.
Fire up the rig (hopefully there will be no smoke) and you should note a dramatic ruction in the strength of the spurs. (Oh yes – you should note the strength of the strongest spurs in your rig before making this change so you can make a before/after comparison.)
Changing the grounding point of the LO cable as noted above also helps (but does not cure) the "W1AW problem". This problem manifests itself by strong signals in the vicinity of 3.58 MHz (W1AW bulletins are transmitted on 3.5815 MHz) being heard over a large portion of the 80M band. According to Index, the problem is caused by a harmonic of the microprocessor 10.24 MHz clock leaking into the LO signal. For the past year they have been using a 10.66 MHz clock which moves the sensitive frequency below the 80M band. Anyone experiencing the "W1AW problem" can contact Index for a replacement 10.66 MHz clock crystal.
There is a noticeable "AGC thump" that can be very annoying on strong signals (both CW and SSB) in the QRP+. The cause of the "thump" is the delay through the switched capacitor audio filter (SCAF) which is inside the AGC loop. The AGC detector is on the output of the filter, and by the time it detects a strong signal and feeds a control voltage back to the IF amplifier, additional components of the strong signal are still making their way through the filter and eventually appear at the output causing the "thump". (This is a rather oversimplified explanation, but hopefully you get the picture.) No amount of diddling with AGC time constants will cure the problem; the only real cure is to go to a dual AGC system or put the gain control element (currently the IF amplifier) after the SCAF. These are major design changes not to be undertaken lightly.
However, there is something that can be done to reduce the effect: limit the signal to a few dB above the AGC threshold until the AGC has time to react to a strong signal. Index has taken this approach in the QRP++ by using a pair of Schottky diodes in the IF to act as signal limiters (D2 and D3 at the output of U4 on the RF board). I tried a similar approach by limiting the audio level into the SCAF. This took care of the thump, but produced a new problem: Strong signals outside the SCAF passband but within the IF passband would activate the limiter causing a sever reduction in the level of the desired signal. I call this an "inverse thump" for lack of a better term. The same problem exists with the IF limiter in the QRP++; it makes no difference whether the limiting occurs at RF or AF. That is the reason for the recommendation to open up the SCAF passband when nearby interference is present (sort of defeats the purpose of a narrow filter...).
The solution that I finally settled on was to place a limiter at the output of the SCAF as shown in Figure 2. This does not completely eliminate the problem, but it reduces the effect from ear splitting "thumps" on very strong signals to just slightly annoying "clicks". When listening to a very strong signal, the best solution is still to switch in the 20 dB input attenuator.
Figure 2. Limiter added to output of U4.
I located the limiter before the second SCAF IC's output buffer amplifier to take advantage of the LP filter present in the buffer amplifier circuit. Five parts need to be added – a 4.7K and a 100K resistor, a 4.7μF capacitor and two diodes – and one existing resistor (R43) replaced. The mod can be made without removing the AF board, but the RF board must be removed. The 4.7K resistor can be inserted by lifting the end of R43 (33K) nearest U4 and soldering one end of the 4.7K resistor into the pad from which one end of R43 was removed. Remove the other end of R43 and insert one lead of the 27K resistor into that pad. The 4.7K and 27K resistors can then be soldered together "in the air" – just make sure that they don't stick up far enough to touch the shield between the RF and AF boards. (Also make sure that the resistor leads don't don't extend below the AF board far enough to touch the shield below!) The positive end of the cap can then be soldered to the resistor junction and the diodes and 100K resistor from the other end of the cap to ground. A convenient grounding point is the ground pad for C20 (the end nearest U4).
An added bonus of this modification is that it also acts as a noise limiter. It will not eliminate noise, but it will keep those static crashes from busting your ear drums!
This modification will probably not help the QRP++. As noted above, these rigs are equipped with an IF limiter. Although that approach introduces a new problem when using narrow SCAF bandwidths when strong adjacent channel interference is present, it has the advantage that less audio distortion is introduced (there is noticeable distortion on the peaks of strong SSB signals using my circuit). I do not recommend that the diodes at the output of U4 be removed since they are also used during transmit and are the basis for the "speech processing" feature of the QRP++. Should you want try this mod with a QRP++, there are some minor differences between the QRP+ and QRP++ that must be taken into consideration:
One of the design compromises in the QRP PLUS is the way selectivity is obtained. The crystal IF filter is basically a sideband filter and the SCAF is used as the final bandwidth determining element. Since the AGC detector is after the SCAF, strong signals that are within the IF filter passband but outside the SCAF passband will be amplified at full IF gain and can cause overloading of the SCAF . This results in various annoying artifacts when trying to copy a weak CW signal and one or more strong signals are close by. Narrowing the SCAF passband does not help, and can even make the situation worse.
The SCAF is implemented using two special ICs, U3 which is a high pass switched capacitance filter and U4 which is a low pass switched capacitance filter. The QRP PLUS SCAF implementation is very similar to one described in QST several years ago [5]. The SCAF ICs contain operational amplifiers that can be used as input/output buffers. These buffers are configured as single pole analog LP filters to eliminate high frequencies outside of the final passband from getting into the SCAF and causing "aliasing" (spurious output response). Ideally, these LP filters should have cutoff frequencies slightly higher than the SCAF upper passband cutoff and a flat response within the SCAF passband. However, this is not possible with simple 1-pole filters, and another design compromise comes into play. The 3 dB rolloff frequencies of these filters are above the highest SCAF passband cutoff, but they produce significant attenuation at lower frequencies. In fact, the high frequency audio response in the QRP+ is determined by these analog filters at the wider SCAF settings. In the QRP++, the LP filter rolloff frequencies have been increased by a factor of about four to improve the high frequency audio response, but they are consequently less effective in attenuating unwanted signals at narrow SCAF bandwidths.
If the analog LP filters before the LP SCAF could be made to have a lower cutoff frequency when the SCAF bandwidth is made narrower, then strong audio signals above the SCAF cutoff frequency would be less likely to overload the SCAF. The modification shown in Figure 3 accomplishes this using diodes to switch additional capacitors into the LP filters. The switching action is controlled by one of the control signals used to set the SCAF HP cutoff (the signal at pin 1 of U3). This results in the capacitors being switched in at SCAF bandwidths of 1.4 kHz or less (1.0 kHz or less if a Rev 03 or earlier EPROM is in use).
Figure 3. Partial schematic of SCAF showing modified
buffer amplifiers and the output limiter.
NOTE: The schematic of the SCAF mod in the original QRP Quarterly article was incorrect!
The original feedback capacitors in all four analog LP filters of the QRP+ should be changed to improve the high frequency response at the wider SCAF settings. In particular, C6, C16 and C18 should be changed to 270pF. I used a value of 680pF for C21 in order to reduce the high frequency distortion introduced by the audio limiter (described above). If you don’t mind a little faster rolloff above 1.8 kHz or so, you can leave C21 at 1000pF. The 1500pF capacitors that are switched into the first three LP filters add substantial attenuation of signals above the narrower SCAF passbands. This greatly reduces the detrimental effects of strong adjacent channel signals on the SCAF response. The capacitors used in the analog filter feedback loops should be good quality, preferably NPO or C0G types.
I mounted the extra caps, resistors and diodes above the board "spider web" style. The junction of the four 10K resistors (I used 1/8W) and the 2.2µF cap (I used a small molded tantalum) is located over the top of U3 Connections to U3 pins 1 and 5 are soldered directly to the IC pins. I mounted a miniature slide switch along the right edge of the board so that I could disable the capacitor switching to facilitate before/after testing, but this is purely optional. A piece of #30 wire-wrap wire snaked along the PC board connects to the switch.
This modification reduces bleed-through and overloading effects in my QRP+ at SCAF bandwidth settings below 1kHz. However, very strong signals within 1.5 kHz or so of the desired signal can still cause problems. The worst problems are from strong signals below the desired signal on 160M though 40M and above the desired signal on the higher bands and within 1.5kHz of the desired signal. About the only way to help this situation – short of adding a narrow IF filter – is to switch in the 20dB input attenuator. (Or add an RF gain control.)
While you have the AF board on your bench, change the coupling capacitor from the SCAF output to the AGC detector (C24 in the QRP+) to 0.47µF in order to improve the AGC low frequency response. Since nothing below about 250Hz gets through the SCAF, there is no need to make this cap any larger than 0.47µF.
You may well decide to forego this modification if your primary mode of operation is SSB. But if you do a lot of CW operating (particularly contests), you should give it serious consideration. If your primarily operating mode is SSB and you have a non-upgraded QRP+, you should at least consider changing C6, C16 and C18 to 270pF (330pF would also work) to improve the high frequency audio response. If you add the audio limiter (and you really should), C21 should not be reduced below about 600 pF.
This modification will have less effect on adjacent channel interference due to the presence of the IF limiter. If you decide to try it, you should note that components are label differently; in particular, the LP analog filter capacitors are C23, C24, C25 and C28. Their values are already 270pF, so it is only necessary to change C28 to 680pF.
Maybe I have been "conditioned" by all the other rigs that I have ever owned, but I really expect the CW sidetone level to be essentially independent of the audio gain setting. Not so with the QRP PLUS! The first thing that I did after installing the mixer protection diodes described above was to see what could be done to correct this situation. I came up with the following fairly easy fix:
That’s all there is to it! The side tone level will still be influenced slightly by the audio gain setting, but not nearly as much as before. I decided that the slight interaction still remaining is OK since I tend to run the audio gain much higher when using the speaker than when using headphones. If I now set the sidetone level so that it is just right when using headphones, it is also about right when using the speaker.
There is a low frequency "thump" when keying the rig that is quite noticeable (and annoying) when wearing headphones. The source of the "thump" is inadequate bypassing of pin 7 of the LM386 audio amplifier. The purpose of the bypass cap on this pin is to isolate the high gain input stage of the LM386 from the power supply. When the rig is keyed, the +12V buss "sags" slightly as the result of the transmitter drawing current. This "sag" produces a transient that results in the audible thump, the amplitude of which depends to some extent on the "stiffness" of the power supply. The problem can be easily fixed by replacing the original 0.1µF bypass cap (C35 in both the QRP+ and QRP++) with a molded tantalum in the range 4.7µF to 10µF having a voltage rating of at least 25V. The positive lead goes to pin 7.
There is also a slight key-up click in the QRP+ that is audible when wearing headphones. This problem is not present in the QRP++. The click is due to a voltage transient induced on C29 when U8 is turned on. The cure is fairly simple [6], but it does require a trace on the PC board to be cut: Insert a 0.47µF capacitor (any value in the range 0.2 – 1.0µF is OK) between R44 and pin 11 of U8, and a 100K resistor from pin 11 of U8 to ground. You can do this as follows (QRP+ only):
More than once I have wondered why a station did not come back to my call only to discover that the RIT was ON and I was calling off-frequency! I therefore decided to add an LED to indicate when the XCVE/RIT/SPLIT switch was not in the XCVE position. The circuit that I used to accomplish this is shown in Figure 4.
Figure 4. RIT/SPLIT ON indicator circuit.
I constructed the circuit on a 0.5 in. x 0.5 in. piece of "perf board" with the resistors mounted upright and the 1µF bypass capacitor hanging below the board. The board is mounted "spider web" style behind the front panel PC board. It must be mounted very close to the front panel PC board so that it does not interfere with removal of the "stacked" PC boards. Input to the circuit is taken from the bottom post of the XCVE/RIT/SPLIT switch. When the switch is in the XCVE position, +5V is fed to the circuit which switches the LED off. When it is in either the RIT or SPLIT position, the LED will be turned on. In addition to the input from the switch, the circuit requires +5V which can be obtained from the back of the front panel PC board at the lower C3 solder pad. The ground connection can be made to the other C3 solder pad, or to the ground plane of the PC board. I used a 3mm "Super Bright" LED mounted in a clear plastic holder. I drilled a hole for the LED holder below the "RIT" and just to the right of "SPLIT" lettering on the front panel.
Being able to control the RF gain would be useful, particularly in the QRP++ with its high IF gain and IF limiter, in reducing adjacent channel interference effects. (My personal opinion is that a variable input attenuator would be the best solution.) Index Labs has provided many QRP++ owners with a simple RF gain control circuit. The circuit, shown in Figure 5, adds a variable DC bias to the AGC line to control the IF amplifier gain.. The circuit connects to pin 4 of U7 (U8 in the QRP+) on the AF board. +V is +10V (+12V in the QRP+) I bread-boarded the circuit in my QRP+ but decided against adding it as a permanent feature. For one thing, there is no place to put the RF gain control except on the rear panel or possibly the top cover.
Figure 5. RF gain control suggested by Index Labs. Added
components are shown in bold.
I found it difficult to read the meter even under good lighting conditions, so I mounted a small "high intensity" green LED behind the meter face. The meter in my rig has an access space behind the face which is translucent, allowing the light to easily shine through. The LED is held in place with a dab of RTV. I believe that several different meter styles have been used in different production runs of the rig, so you may have to devise a different mounting scheme for a light. Since you have to almost completely disassemble the rig to gain access to the meter, you may well decide this is not worth the effort. On the other hand, the LED sure makes the meter easier to read and acts as a convenient power-ON indicator.
To get to the meter, you can remove all four of the boards in the vertical stack. Alternatively, you can leave the XMTR board in place and remove the display board to gain access to the meter. The meter can then be removed by removing the two "L" shaped brackets holding it in place. Once you have access to the meter, you can determine how best to mount a lamp. I highly recommend an LED (green gives a nice friendly glow). The high intensity type will produce plenty of light at a current of 10mA or less. Power for the light can be obtained from the ON/OFF switch, and a convenient grounding point is the ground lug on the bottom of the case. Use an appropriate series resistor for the LED (I used 1.2K) covered with a piece of heat-shrink tubing.
I tried various schemes to illuminate the liquid crystal frequency display, but nothing short of a small lamp directly in front of it was very effective. The display is actually fairly easy to read in all but very poor lighting, so I decided that it wasn’t worth worrying about.
That’s it for my QRP+ receiver mods. The next installment will describe mods for the transmitter. Several mods to improve transmitter spectral purity will be covered, including new output filters for 30M through 10M. So make sure you have plenty of desoldering braid on hand and that your ARCI membership dues are paid so that you won’t miss the next issue of The QRP Quarterly ! By the way – you might want to make sure to have about three T50-10 toroid cores on hand.
My thanks to the many folks with whom I have swapped information on the QRP+ via the Internet, particularly those who tested and provided feedback on some of mods that I have described here.
————
Copyright © 2005, 2012 by Larry East, W1HUE ————
Page last updated on March 22, 2014