RM Italy BLA350V, some mods

The RM Italy BLA350V is a PA designed for HF amateur radio bands, that uses 2 Macom MRF150 RF power MOSfets which deliver an average power of 250W from 1 to 30 MHz. Among the advantages of this amplifier, there are:

• a built-in power supply
• a series of low pass filters, which reduce the emission of harmonics on any operating band;
• various protection systems which give the BLA350V a regular and safe functioning over time.

With good reason, this make the BLA350V a complete, robust and reliable PA unit. However, looking closely at the amplifier circuit, I noticed some imperfections to which I dared to find an alternative. After waiting for the 2-year warranty to expire, I decided to intervene.

CAUTION!

The following modifications have been planned and carried out for the sole and exclusive purpose of optimizing and making the operation of the amplifier more linear, where possible. Once changes have been made, any other feature of the BLA350V (including its RF output power) remain unchanged. Therefore, the Author of this document declines any responsibility should such changes be made for any other purpose.

Furthermore, any modification made to a BLA350V amplifier covered by a post-sale warranty (starting from the removal of the Manufacturer seals) immediately voids the warranty. Therefore, in this case too, the Author of this document declines all responsibility for the execution of the following modifications on any BLA350V linear amplifier for which the Manufacturer’s warranty is in effect.

First of all, the performance of the PA from 160 to 10m, before being modified.

Test 160m

Test 80m

Test 40m

Test 20m

Test 10m

The T2 RF input transformer features:

• a wire type, 3 turns primary winding;
• a pipe type secondary winding consisting of two brass tubes, joined at one end, which make up 1 turn;
• a 3:1 (primary:secondary) turns ratio. Consequently, a 9:1 Zin/Zout ratio (50 Ohm : 5.5 Ohm);
• an Amidon BN43-202 ferrite core (ferrite mix 43).

T2’s role is to match the 50 Ohms, unbalanced input of the BLA350V to the input impedance of the two MRF150 MOSfets (equal to 5.5 Ohms, balanced, approximately). The decision to modify it came after observing that the PA performance was lower on low bands, such as 80 and 160m, where the BLA350V delivered between 200 and 300W. A possible cause of this behaviour could be the core of T2:
if the ferrite permeability was not high enough, its secondary winding (tubular, 1 turn) would not have had enough inductance to couple with the magnetic flux generated by the primary winding and allow for efficient transfer of RF energy, particularly at lower frequencies. The following videos show how T2 behaves between 160 and 10m, on a 5.5 Ohm/4W resistive load.

Test 160m

Test 80m

Test 40m

Test 20m

Test 10m

My mod consists of two impedance transformers, in series with each other, which will be referred to as T2a and T2b. The reasons for this mod is explained in the following description.

T2a

• 3 turns, wire type primary winding, with a 33pF capacitor in parallel;
• 2 turns, wire type secondary winding, with a 100pF capacitor in parallel;
• 3:2 (primary:secondary) turns ratio, resulting in a Zin/Zout ratio 2.25:1 (50 Ohm : 22 Ohm);
• 2 Amidon BN61-202 ferrite cores (ferrite mix 61).

The use of two cores requires the use of longer windings, which gain more inductance, as required on the lower part of the HF. Furthermore, ferrite 61 has a permeability equal to 125, up to 30 MHz. A low permeability allows for windings with 2 or more turns, without the resulting inductance being so high as to cause high SWR on the higher half of the HF, a situation which is difficult to solve with a capacitance in parallel to such a winding. The following videos show T2a’s behaviour from 160 to 10m.

Test 160m

Test 80m

Test 40m

Test 20m

Test 10m

T2b

• 2 turns, wire type primary winding
• 1 turn, wire type secondary winding, with a 820pF capacitor in parallel;
• 2:1 (primary:secondary) turns ratio, resulting in a Zin/Zout ratio 4:1 (22 Ohm : 5.5 Ohm);
• 2 Amidon BN43-202 ferrite cores (ferrite mix 43).

T2b uses two ferrite cores for the same reasons T2a does. Ferrite 43 has a permeability of 850 at 1 MHz, which progressively decreases as frequency increases, reaching 150 (approximately) at 30 MHz. This is useful for the secondary winding of T2b, whose single coil can have enough inductance on low bands. The primary winding of the transformer replicates the secondary winding of T2a, in order to allow optimal coupling with it.
The videos below show the behaviour of T2a and T2b, connected in series with each other, on a 4 resistors load, 22 Ohm/2W each, connected in parallel. During the tests, T2a and T2b are connected in series, but are physically arranged in parallel to each other, to evaluate their performance in the same position in which they would have been mounted inside the BLA350V.

Test 160m

Test 80m

Test 40m

Test 20m

Test 10m

When T2a and T2b are ready:

• remove the 33pF capacitor in parallel with the primary winding of T2a;
• connect the secondary winding of T2a to the primary winding of T2b, and place a 220pF capacitor in parallel to these windings;
• remove the 5.5 Ohm/8W resistive load from the secondary winding of T2b;
• the two core pairs must be arranged in parallel with each other, so that the transformers can be mounted vertically with respect to the BLA350V PCB.

After that:

1. unsolder the terminals of T2’s primary, placing the tip of the soldering iron on each lead pad and carefully remove the ends of the winding from the printed circuit;
2. to unsolder the brass pipes that make up the secondary winding, it is necessary to use a 100W soldering iron (flat tip, screwdriver type) and a 25W iron:
   

   • the 25W soldering iron tip must be applied on the tin that short-circuits the pipes;
   • on T2’s other side, apply the 100W soldering iron tip between the ends of the pipes connected to C14/C15 and C16/C17, to melt the tin on both pads;

   

   • with the utmost care, remove T2 from the PCB;
   • clean the pads with the soldering iron, and remove any residual tin from the PCB.
3. remove C12 from the PCB, with the same technique used at point 1, and mount a 22pF capacitor in its place;
4. solder a 820pF capacitor on the pads the secondary of T2 was soldered to;
5. mount T2a and T2b in place of T2:
   • the primary winding of T2a (3 turns) must be soldered to the lead pads the primary of T2 was soldered to;
   • the secondary winding of T2b (1 turn) must be soldered to the T2 tubes’ pads;
6. check that everything has been done correctly and without damage.

When the mod is complete, the new T2 should look like this:

The C14/C15 and C16/C17 capacitor pairs, placed between the secondary winding of T2 and the two Macom MRF150 MOSfets, prevent the bias voltage (Vgs) from one of them to reach the other transistor through the winding. Since an MRF150 may require a Vgs which is different from another MRF150, it is necessary to prevent these voltages from overlapping. Each pair has a capacitance of 4.4nF. Although it seems a value chosen to implement a BLA350V gain control mechanism, raising this capacity allows the RF to reach the two MOSfets more easily, especially at frequencies below 7 MHz, improving the PA performance on low bands. At the same time, the two Vgs remain isolated from each other. I applied a 100nF ceramic capacitor in parallel to each pair: the following photo illustrates the mod.

After the C14/C15 and C16/C17 pairs, are the gates of the two Macom MRF150s. This is the endpoint for the RF that crosses these capacitors and the bias voltage (Vgs) for each transistor. The only part that refers these voltages to ground is the input impedance of each MOSfet: but the input impedance varies like the RF applied to the transistors, as does their gain. The use of a resistor (Rgs) between the gate of each MRF150 and ground introduces a reference whose value is independent of the RF and makes the amplifier's operation more linear. Each Rgs must have a sufficiently high value to avoid being a loss element for the RF and the bias voltage that arrive at each MOSfet: taking the Motorola EB104 as a reference, I used two 10 kOhm/¼ W resistors and I soldered one of them between the gate and one of the two sources of each MRF150, as the following photo shows.

T4 is the RF transformer used to couple the output of the two MRF150 MOSfets to the input of the active low pass filter within the BLA350V. Its primary winding is made with two small brass tubes, short-circuited at one end, which make up 1 turn; its secondary winding is a 2 turns, insulated wire winding. Since the primary: secondary turns ratio is 1: 2, the impedance ratio is 1: 4 (its square):

• input impedance 12.5 Ohm (the drain-to-drain impedance of the MRF150s;
• output impedance 50 Ohm (the input impedance of each low pass filter).

A binocular ferrite core (multi-aperture core, also called pig-nose core) was used for T4: I do not know which ferrite mix was used to produce this core, but it is realistic to think that it is a BN43-7051 core (mix 43). It is enough to approach it to the RF output transformer of my TS-140S (Pout = 100W max.) to realize that T4 is abundantly undersized and unable to accommodate the magnetic flux generated by 300W of RF without having unnecessary and harmful losses. First of all, the overheating of the RF output transformer in case of a prolonged use of the BLA350V. Overheating means that it is impossible to touch T4 with your hand. This means that the core of T4 must withstand an amount of RF power higher than its specifications, which represents a waste of RF energy and a potential danger to the integrity of T4 (from both an electromagnetic and physical point of view). The only alternative is to replace T4 with a transformer that has similar characteristics, but is sized to withstand the power involved.

The schematic diagram of the BLA350V tells that the primary winding of T4 is connected to a balanced RF generator (the output of the push-pull RF power stage of the PA), while its secondary winding is connected to an unbalanced load (the low pass filter input). In addition, the primary of T4 does not have the center tap, the intermediate socket that is used to DC saupply the two MRF150. In this situation, the best replacement for T4 is a balun. Specifically, a balun that has one side with a 12.5 Ohm balanced impedance, and the other side at 50 Ohm unbalanced. The W2FMI-4: 1-HB50 balun, realized by Jerry Sevick (W2FMI, SK) and presented in the book "Transmission Line Transformers Handbook" published by Amidon, has these characteristics:

• it is a balun which derives from the Guanella balun schematic (current balun);
• it does not use primary and secondary windings, but transmission lines appropriately connected to each other. Therefore, it has a higher bandwidth and a constant Zin/Zout ratio compared to a traditional transformer (based on inductive coupling);
• each line was wound on an Amidon R61-037-300 ferrite rod;
• the balun withstands up to 1kW of RF power, and has an efficiency of 99%.

After receiving the necessary components to homebrew the W2FMI-4: 1-HB50, I followed Dr. Sevick’s instructions and I tested my DIY balun on a resistive load (10 resistors, 12 Ohm/2W each, in parallel to one another). The result confirmed the characteristics of the balun, and can be seen in the following videos.

Balun W2FMI-4:1-HB50
Test conditions

Balun W2FMI-4:1-HB50
160m test

Balun W2FMI-4:1-HB50
20m test

Balun W2FMI-4:1-HB50
10m test

Replacing T4 with the 4: 1 balun can be done by overlapping. That is, the input terminals of the balun can be soldered to the pads to which T4’s are soldered; the same applies to the output terminals. No other modifications to the BLA350V circuit are required. After the removal of the amplifier chassis, the replacement steps are as follows:

1. it is necessary to unsolder the terminals of the secondary winding of T4. It is not necessary to remove the entire PCB from the chassis: just apply the tip of the soldering iron on the pad to which each of these terminals is soldered, until the solder melts. Now, carefully remove the end of the winding from the PCB. After having unsoldered both ends, the first step is performed.
2. it is necessary to unsolder the brass tubes that make up the primary winding of T4. The best way to do this is to use two soldering irons, 100W each (flat tip, screwdriver type):
   1. apply the tip of one iron on the tin which short-circuits the pipes;
   2. on the other side of T4, apply the tip of the second iron between the ends of the tubes connected to C21 and C22, so that it can simultaneously melt the tin on both pads;
   3. with the utmost care to avoid damaging T4, its pads, and the rest of the PCB, remove T4 from the BLA350V circuit;
   4. clean each pad with the soldering iron, leveling any vertical tin protrusions, and make sure that there is no free tin (little balls, filaments, etc.) remaining on the PCB.
3. 7. Now, the W2FMI-4: 1-HB50 balun can be mounted in place of T4. Compared to it, the balun has more demanding dimensions. Most likely, it will be necessary to leave it a few centimeters above the PCB, and to use enamelled copper wire “extensions” to connect its terminals to the BLA350V circuit. In any case:
   • on the side where the transmission lines are connected in series with each other (50 Ohm unbalanced side), the two free ends can be connected to the pads used by the primary winding of T4 (the wire type winding);
   • on the side where the transmission lines are connected in parallel with each other (12.5 Ohm balanced side), the two ends of this parallel can be connected to the pads used by the primary winding of T4 (the tubular type winding ).
4. After the replacement, check that everything has been done correctly, without any damage, and that there is no object (tin or else) free to move on the PCB.

Before testing the modified BLA350V, two other small mods are necessary: after exiting T4 (now, after the balun), the RF passes through two toroids, transformers T5 and T6, but the conductor carrying the RF inside them is undersized compared to the 300W supplied by the PA. The disproportion is evident by comparing its diameter with the diameter of the enameled copper with which the toroids in the amplifier's low pass filters are wound. Within T5 and T6, there is a small brass tube soldered (at one end) to the PCB of the BLA350V: for both toroids, the conductor to be replaced is inside the tube, and must be unsoldered/removed with the utmost care and maximum patience. The new conductor (an enameled copper wire) must have a diameter equal to that of the balun’s transmission lines, or not less than that of the enameled copper in the amplifier output filters.

The most obnoxious mod. Given their size and the way they were installed, T2a, T2b and the balun that replaced T4 require the space occupied by the fan on the top panel of the chassis. In order to close it properly, it is necessary to move the fan above the panel, so that it remains outside the BLA350V. The aesthetics of the linear is much better if the fan remains inside, but if the amplifier is located on the highest point of the shack, to properly dissipate the heat generated, this change can also go unnoticed.

Testing the BLA350V must be performed on a dummy load connected to the PA output, and should verify that the RF output power is not less than 250W. I have taken the tests performed with my BLA350V, and the videos are below.

BLA350V
Test conditions

Test 160m

Test 80m

Test 40m

Test 20m

Test 10m

If you want, after testing the amplifier on each operating band, there is another particular test. With the BLA350V in stand-by mode, try to touch the balun that replaced T4: the conductors of the two transmission lines can burn, the ferrite rods are warm.

A Macom MRF150 is an RF power transistor with a 17dB gain. That is, the output power is 50 times (approximately) the input power. All the functional tests I made with my BLA350V were performed with a 10W driving power. Which is halved in the presence of the attenuator ATT1. Therefore, the two MRF150s get 5W each, and 250W come out from these MOSfets. This is the nominal power of the BLA350V: any other measurement is the result of harmonics and spurious signals that add to the fundamental RF signal exiting the amplifier.
As mentioned at the beginning of this document, the mods described were not made to increase the performance of the BLA350V, but to eliminate some inefficiencies that I found and to optimize its operation (to make it even more linear, if possible). At this point, it is perfectly normal that:

• the RF output power has not increased, compared to the values measured before the mods. Even better if it has decreased: it means that the emission of harmonics has decreased too, and that the amplifier works in an even more linear way;
• the difference, in terms of RF output power, between one operating band and the next (or previous) band has decreased. In addition to the increase in linearity, this is due to the fact that the gain of the amplifier has stabilized over the HF range.

However, there is one last mod which I did not make due to lack of the necessary components: the retouching of the output low pass filters. This modification, if done correctly, would reduce the RF power delivered by the BLA350V on many bands, eliminating the harmonics underlying many false RF output power readings, from 345W on 10m to 450W on 40m. And by leveling the amplification to the 250-300W that this excellent PA can produce with an input equal to 10W or little more.