Fullscreen HTML5Med Res (??MB)HTML4

ASK SILICON CHIP Got a technical problem? Can’t understand a piece of jargon or some technical principle? Drop us a line and we’ll answer your question. Send your email to silicon<at>siliconchip.com.au Tapped Horn Subwoofer orientation I have built the Tapped Horn Subwoofer (September 2021; siliconchip. com.au/Article/15028) but I cannot figure out which part is supposed to be the top. In the article, the author has possibly swapped the top and bottom designation in the part about the fitting of the final panels. I tried to look carefully at the author’s photos in the construction article, but there is no photo of the finished speaker looking at the front panel, with the exit port visible. It probably does not matter, but my question is: should the exit port be at the top or close to the floor? Regardless, it is an excellent article/idea/construction project. (P. M., Loftus, NSW) • Phil Prosser responds: I put the “mouth” of the horn at the bottom. Ideally, it should face into the corner of the room. My reasoning is that the base is screwed on but not glued, allowing access to the driver, so it will not have screw holes neatly filled and invisible. Also, I think that having the port right at floor level in the corner will more effectively couple it into the room. Though given the frequencies involved, this is almost certainly of negligible impact. As I noted in the text, a tapped horn is not horn loaded in the way the name suggests and is more like a resonant system with many parallels to a transmission line. To this end, the relationship of the exit from a tapped horn to the walls/floor is less critical than with a conventional horn. I did model the tapped horn in ‘quarter space’ (into a corner) and ‘half space’ (against a wall) with a preference for the corner, as that is a good place to hide this thing in a typical house. This placement does not substantially impact the roll-off rate like with a regular horn, where you can create an effective extension of the horn using reflections from the walls. Beyond this, I do not have an opinion on up or down. I hope people try 92 Silicon Chip these things out and get into using the “Hornresp” software; have a play and share your experience. Battery Balancer with a stack of LiFePO4 cells Regarding your High Current Battery Balancer projects (March & April 2021; siliconchip.com.au/Series/358), does it redirect charge to cells that are discharging more quickly than others in the stack, or is it only active when charging? Eg, see siliconchip.com.au/ link/abbb I’m tired of replacing SLAs in my UPS. I want to replace them with an 8S1P LiFePO4 battery – and this project looks like it has all the right features, but there is the issue of the behaviour of the built-in charger. What can I do about this? I’m also keen to apply a Balancer to a 15 cell stack of LiFePO4 cells, notionally rated at 150Ah charged by a 40A charger. Is it possible to daisy-chain four balancers to achieve this, or is the limit of two a hard limit? As always, a great magazine! (M. S., Doncaster East, Vic) • The Balancer is always active as long as the cells are between the low and high voltage settings. So, provided the battery is not over-discharged (or under extreme load), the Balancer will transfer charge from cells with higher voltages to those with lower voltages (as long as the difference is larger than the threshold setting). The Balancer is entirely independent of the charger: its only job is to keep the cells in balance. It aims to be invisible to the charger. So in that regard, you cannot change the battery chemistry to a type not supported by (or suitable for use with) the charger. There are numerous LiFePO4 rechargeable batteries that are designed as drop-in replacements for lead-acid types, so those are probably your best bet. You could theoretically charge a Li-ion/LiPo pack safely with a multistage lead-acid charger. It would need Australia’s electronics magazine enough cells that its fully charged voltage is higher than the termination voltage of that charger, with cells that can handle the peak charging current. However, you would almost certainly be leaving a significant percentage of the battery capacity unused if you did so. For example, if the charger delivered a maximum of 10A and terminated at 14.4V, you could use four Li-ion/LiPo cells in series as long as they can handle the 10A charge rate. But that battery would normally have a fully charged voltage of around 16.8V, so you would only be charging them to about 50% of their full capacity. LiFePO4 is better suited to the role as 14.4V = 4 × 3.6V. Currently, the Balancer is only designed to operate by itself or in pairs. It would be possible to construct a separate device to allow more than two units to be used together, but we have not done that yet. We will consider doing so. One thing that might be possible, but we have not tried, is to arrange two Balancers in pairs, balancing eight cells per pair, then connect the bottom eight of your 15 cells to one Balancer and the top eight to the other. There would be one cell common to both Balancers, and therefore (in theory at least), all 15 cells would ultimately become balanced. However, it might not work efficiently, and the common cell might wear out faster than the others as all balancing current between the top and bottom halves would flow through it. Battery Multi Logger questions & suggestions This project (February & March 2021; siliconchip.com.au/Series/355) is welcome and looks like an extremely useful addition to any type of off-grid service battery installation. I’m mainly thinking of batteries for UPSs, remote sites or caravans. I have some questions regarding the operation. How many High Current Battery siliconchip.com.au Balancer modules are supported by the Battery Multi Logger? I suspect that only one is supported as there is no discussion of this capability. Suppose only one High Current Battery Balancer is supported. In that case, one might think of using “nested” Balancers: four Balancers, each balancing a stack of four cells and a fifth Balancer monitoring the four stacks connected to the Battery Multi Logger. Is this a possible solution? Is the Battery Multi-Logger data logging retained as originally described, ie, hours/days/weeks? I’m a little disappointed that the cell voltage history seems to be limited to “around 100 data points” to give “around 15 minutes of balancing data”. I think there is a way to increase the “history” to a much larger number as cell balancing for large capacity LFP batteries/cells can take a long time. A 32Mb SPI flash memory chip is under $10. (M. S., via email) • You’re right in that the Manager can only communicate with one Balancer directly. The design is quite squeezed for space and I/O ports, so it would not be an easy task to add more serial ports to it. The nested solution would likely work but requires quite a bit of extra hardware. The data retained by the Manager is less than the Logger due to the need to set aside space for storing the Manager settings for the Balancer and Soft Switch in the Micromite’s limited “VAR SAVE” space. Using an external flash chip makes sense, but that would require a fair bit of the program to be reworked. Since it is already quite close to reaching the flash memory limit, adding this extra capability wouldn’t be easy, even with storage offloaded to the external chip. We’ve run into similar problems with many 28-pin Micromite-based projects; it doesn’t take much code to occupy all the flash space (and RAM too). So expanding its storage capabilities will probably involve entirely new hardware. Perhaps a better option would be to regularly send the data to another device over the serial port (especially as functions already exist to do this). There is a function that is called each day to update the stored data; it could be done then. The receiver could be another micro that simply receives and logs data to siliconchip.com.au an SD card or flash chip. However, this will probably affect the low-power performance of the Logger/Manager. Reducing soldering iron power I am using an old-fashioned soldering iron to push brass inserts into 3D-printed plastic. As the full bore temperature is a bit on the hot side, I need a means of reducing the temperature. The actual temperature is not very critical. I am looking at some sort of PWM Mosfet arrangement with a knob to twiddle the temp. Do you have anything in your catalog which would do this? (G. C., Mount Dandenong, Vic) • Presumably, this soldering iron is mains-powered, in which case a phase controller would be suitable. The Refined Full-Wave Motor Speed Controller (April 2021; siliconchip.com. au/Article/14814) would be suitable, although perhaps more complicated than necessary. The feedback control is not required. Depending on the wattage, a standard light dimmer would also work if installed in a suitable enclosure with mains input and GPO (mains socket) for the soldering iron. Our Heat Controller design is also suitable (July 1998; siliconchip.com. au/Article/4687), and the PCB for that project is still available from our Online Shop. Original Colour Maximite limitations I cannot find any reference on TheBackShed forum to the Maximite computers. I need to produce a completely random dice roll using MMBasic for something I am working on using the original Colour Maximite. The other program I want to do is for resistor colour codes, but it does not have the right colours. Is there any way around that besides using the Colour Maximite 2? (R. M., Melville, WA) • There is some discussion of the Maxmite computers on TheBackShed forum, but it doesn’t seem to be organised into any specific location. You have to search the forum. For random numbers, see the RANDOMIZE and RND functions in the MMBasic Language Manual Ver 4.5 published for the original Colour Maximite. Australia’s electronics magazine The Colour Maximite only supports eight colours as the hardware only has three digital lines physically driving the VGA connector. Since there are ten resistor colour codes and the Colour Maximite only supports eight colours, there is no way it can produce them. You definitely need something like the Colour Maximite 2. Using BMP280 sensor with the Micromite Many of your projects have incorporated the BMP180 pressure sensor (eg, the Micromite LCD BackPack V3 from August 2019; siliconchip.com. au/Article/11764). The BMP280 has superseded it, but I don’t think the code for the BMP180 is still valid for this newer device. Would you consider doing an El Cheapo project on using this device with a Micromite? (J. H., Nathan, Qld) • We mentioned the BMP280 in the December 2017 El Cheapo Modules article on page 82 (siliconchip.com. au/Article/10909); however, we did not provide any example code for it in that article. The main difference is that it offers higher resolution readings. You can find Micromite software for interfacing with a BME280 at www. thebackshed.com/forum/ViewTopic. php?TID=8362 The only difference between the BMP280 and the more expensive BME280 is that the latter incorporates a humidity sensor. If you remove all the code from that example that has to do with the humidity measurements, it will work with a BMP280 instead. Improving GPS Disciplined Oscillator I just completed building your Programmable GPS Synched Frequency Reference (October & November 2018; siliconchip.com.au/Series/326) and it works very well. However, there is 0.8V of ripple on the output at about 116MHz, which appears to be from the 3.3V supply. The ripple on the 3.3V supply is about 0.1V near the DAC chips and about 0.8V near IC4, but I can’t tell where it’s coming from. Also, the PLL chip (IC2) gets very warm. The Micromite by itself does not have this ripple, and I can’t see any large oscillations on the unused outputs on IC4. I tried adding a ground plane to November 2021  93 the underside of the board and that improved things slightly. It appears the signal I am seeing is from approximately 100MHz ringing on the Schmitt trigger outputs (IC4 & IC5). I have seen this mentioned elsewhere on the internet and in TI’s application notes. I confirmed this by removing IC5. I note you have included damping resistors on the outputs. Any suggestions to reduce this ringing? I added a wire from pin 7 of IC5 to the ground side of the bypass capacitor but it made little difference. I also piggybacked a 10nF and 10μF capacitor onto the 100nF capacitor, but that too made little difference. I ran a 3.3V and a ground wire directly from the header connector to the connections near IC5 on the board’s underside, which reduced the ringing a bit. I also added the extra ground links per your previous suggestions and my ground plane on top of that. That made the unit much more stable but still with some ringing. I am also finding that at 80MHz+, the signal from the PLL is barely enough to trigger IC5. I tried 22pF and 4.7pF capacitors across the 510W resistor, but they made the PLL unstable. I then paralleled a 560W resistor with the 510W resistor and that seemed to fix the problem, but I am not sure if that is a good idea. The only remaining problem now is that the outputs of IC5 are sensitive to what is happening at its other inputs/outputs. Changing the settings for CON4 affects the signal from CON3. However, the output from CON3 (IC4) is now very stable. I have the unit in a box now, and with the extra ground links, ground plane and 560W parallel resistors, it is working well. There is still a bit of ringing on the outputs but not too bad. The GPS receiver (VK2828U7G5LF) fits perfectly into a 2xAA battery box with room for a magnet for mounting it on the top of something steel. I also put a piece of clear perspex wrapped in reflective tape on the bottom, poking through a hole in the case in an attempt to make a light pipe to see the flashing green LED. To my surprise, it worked pretty well! Finally, there is something strange happening during GPS disciplining. The c-value often jumps to 0 or 16,777,215 on an update and is rarely in between. Any ideas what could be 94 Silicon Chip causing this? (M. H., Mordialloc, Vic) • You appear to have seen the previous suggestions we published from readers to attach extra ground wires in parallel with the PCB traces to decrease the impedance for the ground return currents. Another reader suggested adding more bypass capacitance around the PCB. Extra wire links could also be added to the 3.3V rail. The 510W resistors simply limit the current out of the PLL pins; there is certainly some scope to reduce their value. At those frequencies, the input pin capacitance is probably the next most significant load; we estimate around 400W at 100MHz. So your parallel 560W resistors should be fine. With IC5 appearing to be the culprit, you could also look at adding extra supply bypassing to it. I would also have a close look at its pins. We can’t see any errors in the photos you sent, but we have been caught more than once by a pin that appeared to be soldered correctly but wasn’t. It’s good to hear that the extra grounding helped. It sounds like the disciplining code is overreacting. Try reducing the Gain Value as discussed on p85 of the second article and see if that stops the overshooting. We wonder if the micro is not detecting or counting the frequency correctly, which would also cause erratic behaviour. How does the 40MHz signal look, from pin 19 of IC2 through LK1 to the Micromite TX pin? Choosing a solar panel to charge a battery Thank you for your wonderful magazine. I’ve enjoyed assembling projects going back to Radio Television & Hobbies and have put together some 30-40 kits. I’m not good at theory but can follow construction guides to successfully complete most projects. Recently I built the 12/24V 3-Stage MPPT Solar Charge Controller Rev.1 (March 2012; siliconchip.com.au/ Series/29) but have not yet bought a matching solar panel. I have two Mobility Scooters and two Mobility Wheelchairs in my household; each is powered by two 12V gel batteries connected in series for 24V. All batteries are identical Sonnenschein GF1244Y rated at 44Ah (C5) or 50Ah (C20). The batteries in the two wheelchairs have just been replaced with new, identical batteries. Australia’s electronics magazine I intend to keep charged the old batteries as two separate 24V systems because they still have a useful life for other purposes. The solar panel I am considering is as follows: Maximum Power (Pmax): 250W Voltage at Maximum Power Point (Vmp): 37.8V Current at Maximum Power Point (Imp): 6.6A Open Circuit Voltage (Voc): 44.8V Short Circuit Current (Isc): 6.9A Type: Q-Cells Grade A Monocrystalline Maximum System Voltage: 1000V DC Operating Module Temperature: -45°C to +90°C These are, in some aspects, a little above those recommended in the original article from February 2011 (pages 40 & 41). Is this panel a satisfactory match, or should I look for another? (K. U., Sunbury, Vic) • The solar panel you have chosen is well-suited to the lead-acid gel-cell batteries you are using. Wideband Oxygen Sensor revisions I bought a Jaycar KC5485 Wideband Oxygen Sensor Controller kit. I want to use a Bosch 4.9 O2 sensor with the Bosch CJ125 control chip. Can I integrate that into your kit? Is there any upgrade? (T. M., via email) • The KC5485 kit is based on our original Wideband Oxygen Controller design from the September and October 2009 issues. That was superseded by an improved design (June-August 2012; siliconchip.com.au/Series/23). Both used the Bosch 4.2 wideband sensor with a microcontroller to control the sensor. They do not support the Bosch 4.9 sensor, which would require a new design. Both units provide a narrow band output to simulate a narrowband sensor for the engine so that you can replace the old one entirely. That way, the wideband sensor can be installed near the engine and accurate Lamdba readings taken with the engine controlled via the narrowband simulator output. As we are using a microcontroller in both cases, we are not using the Bosch CJ125 controller chip. It is not possible to integrate the CJ125 on either of the Wideband Controller boards we published. continued on page 96 siliconchip.com.au Advertising Index Altronics.................................69-76 Ampec Technologies.................. 13 ADI Maxim Integrated................. 11 Dave Thompson.......................... 95 Dick Smith Contest..................... 77 Digi-Key Electronics...................... 5 Emona Instruments.................. IBC Hare & Forbes............................ 2-3 Jaycar.............................. IFC,45-52 Keith Rippon Kit Assembly......... 95 LD Electronics............................. 95 LEDsales..................................... 95 Microchip Technology......... 7, OBC Mouser Electronics....................... 9 Ocean Controls........................... 10 Phipps Electronics...................... 85 PMD Way..................................... 95 SC Christmas Decorations......... 63 Silicon Chip Shop.................90-91 Silicon Chip Subscriptions........ 68 Solder Master............................. 15 Switchmode Power Supplies..... 12 The Loudspeaker Kit.com.......... 87 Tronixlabs.................................... 95 Vintage Radio Repairs................ 95 Wagner Electronics..................... 14 Next issue release The December 2021 issue is due on sale in newsagents by Thursday, November 25th. Expect postal delivery of subscription copies in Australia between November 23rd and December 13th. 96 Silicon Chip Controlling motor speed with an Arduino I want to build your June 2011 Motor Speed Controller (siliconchip.com.au/ Article/1035) using the Jaycar KC5502 kit. Can the external potentiometer be replaced by connecting to an Arduino for speed control? I have been purchasing motor controller boards from the website: www. dimensionengineering.com (Sabertooth 2x12 R/C boards). This alternative would be of great interest to my group of friends that also build Star Wars robots. I am scratch-building robots having left aircraft engineering after 36 years. One of my builds was in DIYODE magazine issue 3. (F. H., Engadine, NSW) • You can use an Arduino to control the June 2011 Motor Speed Controller via a pulse width modulated (PWM) output. The PWM signal is filtered to give a varying voltage to feed the Speed Controller. Disconnect the trimpot or potentiometer connected to the +IN1 input (pin 1) of IC1 and instead, connect the PWM output of the Arduino via a 10kW resistor. The supply ground of the Arduino must connect to the Motor Controller ground. PWM frequency should be set for at least 500Hz so that the 10kW resistor and existing 1μF capacitor at the +IN1 input provide a smooth DC voltage. The motor speed can then adjusted by varying the duty cycle of the PWM signal from the Arduino. A 50% duty will result in 2.5V DC at the Motor Controller input and will set the motor at full speed. A 0% duty cycle will turn the motor off. If you want the motor to run at maximum speed when the duty is set at 100%, change the 10kW resistor to 22kW and connect another 22kW resistor from the +IN1 input to ground. This will effectively halve the applied voltage. Details on using PWM with Arduino is covered in the tutorial at siliconchip. com.au/link/abb7 Transformer for 20W Class-A Amplifier Where can I obtain the 160VA 16-016V shielded toroidal power transformer for the 20W Class-A Amplifier Module project (May-September 2007; siliconchip.com.au/Series/58)? Australia’s electronics magazine In Leo’s April 2011 article titled “Fixing transformer buzz in the Class A amplifier”, he suggests using an 18-0-18V transformer when the mains is reduced to 230V AC (it’s around 239V AC here in Melbourne suburbia) and adding chokes which reduce the DC output voltage of the power supply by more than a volt. As ordinary 16-0-16V toroidal transformers are not off-the-shelf items (it’s either 15-0-15V or 18-0-18V), should I be looking to get an 18-0-18V transformer? In the 1998 15W Class-A amplifier version, its designers had lots of trouble with hum, even when using the transformer inside a steel box inside the amplifier chassis. Would an ordinary unshielded 160VA toroidal transformer be OK if placed about 400mm below the Class A amplifier PCB, as in the separate power supply version of 1998? (I. H., Essendon, Vic) • We checked all four articles and couldn’t find any mention of a source for the transformer. These were supplied with the contemporary kits but they have long since sold out. It was probably made by either Harbuch or Tortech. Regardless, either of them should be able to wind one for you. Email sales<at>harbuch.com.au or see siliconchip.com.au/link/abba Don’t forget to specify the electrostatic shielding. You should be able to use the 18-018V transformer. If the voltage ends up being too high, wind on extra turns in the opposing direction to the original winding for each 18V winding to reduce the voltage. But the chokes will probably drop the voltages enough. Placing the transformer in a separate box will reduce the hum significantly. 400mm spacing from the amplifier should be sufficient. If hum is still heard, try rotating the toroid to get the lowest hum or move the toroid further away. Finally, note that the 20W Class-A amplifier is essentially made obsolete by the Ultra-LD Mk.3 (July-September 2011; siliconchip.com.au/Series/286) and Mk.4 amplifiers (August-October 2015; siliconchip.com.au/Series/289), which have similar distortion (lower in the case of the Mk.4), significantly more power and higher efficiency. Importantly, they also do not suffer from the same hum problems that plague Class-A amplifiers. SC siliconchip.com.au