The Beast Within

We finally received our Happy Hacking Keyboard (HHKB) Pro 2 that we ordered for our iPad. The idea was to relegate our Apple Wireless Keyboard to only those trips where we need extreme portability and still require a keyboard. The rest of the time, we’d use the HHKB because of its superior tactile feel. As the HHKB Pro 2 is wired, we spent a little time trying to determine the difference in power draw between it and the Apple Wireless Keyboard, to see just why reports suggest the HHKB Pro 2 doesn’t work with the iPad, and how our iPad’s battery life would change if we did get the HHKB Pro 2to work. It turns out, finding a good answer is difficult, because Apple doesn’t seem to make the power consumption specifications of its accessories easy to find, if it makes them available to the public at all.

There’s power draw at both ends.

When we’re talking Bluetooth, we’re talking power draw on both the device end (because of the Bluetooth antenna), as well as on the accessory end. The former is why most Bluetooth accessories have batteries. The power draw here is more measurable, and we can estimate it fairly.

For example, one third-party wireless keyboard manufacturer advertised their keyboard as drawing 6 mA of power, which serves as a reasonable estimate for what the Apple Wireless Keyboard draws.

Something to keep in mind is that the iPhone 4S and new iPad are using the latest Bluetooth revision (4.0), which, among other things, is designed for lower power consumption for accessories. Depending on the source, reports suggest Bluetooth 4.0 cuts power consumption by 30-50% compared to earlier Bluetooth revisions. A lot of this is dependant on how frequently a given accessory communicates with its host, as the low power consumption of Bluetooth 4.0 seems linked to an accessory’s sleep mode, wherein it leeches only tens of nA.

During peak transmission/reception, however, an accessory even under Bluetooth 4.0 is using a lot more power, at tens of mA. Of course, this all assumes the accessory is Bluetooth 4.0-compliant, which few yet are. Certainly, our Apple Wireless Keyboard is not, so we’ll continue to assume it uses roughly 6 mA of power when in use.

Back to the antenna itself, despite the fact most “This is what you can do to conserve iPhone/iPad battery power” guides claim you should turn Bluetooth off to maximize your iDevice’s power, the reality is the difference between keeping it on and turning it off is very small. This may be due to the Bluetooth antenna being part of a larger RF board which shares resources among different transmission protocols, but because Bluetooth is already very power efficient on the device end, the additional draw here is small. Several folks online have graphed power use between smart phones with Bluetooth on vs. Bluetooth off, and the difference really is really negligible. There’s no reason to assume the iPhone would operate much differently in this regard.

Even if there were a notable difference, it would only matter to you if don’t use Bluetooth with other accessories and only with a wireless keyboard; if you need to keep Bluetooth on regardless of what kind of keyboard you use, the power consumption here doesn’t matter, unless there’s a substantial change in consumption the more accessories are connected, or the more the keyboard is used.¹

In our case, we primarily use Bluetooth for our keyboard, and occassionally use it for our Jawbone Jambox. So technically, when we’re not jamming out to music, perhaps our current setup is burning more power than 6 mA after all, but even if we rounded up to 10 mA, it ultimately doesn’t matter when we consider the HHKB Pro 2.

Bad news for the HHKB.

Per the HHKB Pro 2 specs (which were conveniently a lot easier to find), the keyboard draws a max of 100 mA, and 30 mA during normal operation. The max draw is a lot higher because it assumes you’re using the built-in USB hub and have other devices connected to it. With no devices attached, the keyboard’s draw is 30 mA, which is about five times more power than what the Apple Wireless Keyboard draws on the keyboard side.

With iOS 4.2, Apple reduced the power consumption allowance through the camera connection kit (CCK) from 100 mA to 20 mA. This explains why the HHKB Pro 2 won’t power off the CCK anymore without auxiliary power, and why other USB keyboards get adequate power through the CCK to function. We can roughly estimate that a USB keyboard uses roughly three times the mA a Bluetooth keyboard uses.²

Our solution is to use auxiliary power, but of course this means we need to have an outlet handy, which typically limits us to buildings. We explored the idea of disabling the HHKB’s onboard USB hub, thereby disabling the additional power draw, but it seems this requires more serious hardware tinkering. As the HHKB is not a cheap keyboard, we decided not to risk damaging anything, so instead explored other routes.

Typically, a USB keyboard offers no need to worry about extra batteries, which is always a potential concern when traveling with a Bluetooth keyboard. For us, the determination as to whether we use the HHKB Pro 2 nearly exclusively with the iPad comes down to whether or not we can reduce the power consumption by at least 10 mA. Otherwise, we’ll still have to keep the Apple Wireless Keyboard around for when we don’t have an outlet handy.

Using a powered-hub.

When we first received the HHKB Pro 2, we of course had to verify that it wouldn’t work with the iPad as-is. That was quick, and while expected, still disappointing.

We had ordered several USB widgets to experiment with, but as we hadn’t received them yet when the HHKB arrived, we were limited to playing around with what we did have on hand: a USB hub.

The USB hub we pulled from our cable box is a small four-port powered hub made by IOGEAR. We bought it several years ago for use with, we believe, an Apple Powerbook or Macbook, and it’s collected dust ever since. It’s a 2.0 USB “microhub” with a USB cable, approximately 2″ in length, attached.

Our initial test was with the USB hub unpowered, power cable unattached. We plugged the HHKB into the hub, the hub into the CCK, and the CCK into the iPad. Unlike our attempt without the hub, we actually received feedback from the iPad this time, with the iPad complaining that there wasn’t enough power to use the attached accessory. We don’t know why we didn’t get a similar error when we didn’t use the hub, but at least we got some feedback. Unfortunately, the keyboard still didn’t work.

That was easily fixed when we plugged the power cord into the wall and hub, however. The HHKB sprung to life, and defaulted to the Japanese keyboard. We removed that setting on the iPad (leaving just English and Emoji keyboard profiles installed), and we were off.

While this was a great start, it offers marginal benefits, as plenty of tactile keyboards can power directly off the CCK. The reason the HHKB Pro 2 grabbed our attention is because of its size, and while this remains beneficial at home because setting our portable desk up anywhere is easier, the HHKB Pro 2′s small footprint can best be leveraged if we can turn this setup into one that’s more portable. Another problem with this setup is that the iPad cannot be charged when the keyboard is attached. Even though the USB hub is powered, it doesn’t supply the iPad itself with any power.

Streamling the setup.

There aren’t many keyboard aficionados trying to get power-hungry tactile keyboards working with the iPad, but there are a good number of audiophiles trying to get DACs working with the iPad. So when perusing for DAC solutions, we came across the suggestion of using a USB Y-cable.

The one we picked up had USB-A on one end, and on the other end USB-A and mini-USB-A. This type of cable is typically used to power USB hard-drives, using the extra USB-A end to obtain additional power from an unused USB port on a computer.

For our experiment, we plugged the end with only one cable into the CCK, and plugged the mini-USB-A cable into the HHKB. The other USB-A cable we plugged into the iPad power brick, which was plugged into an outlet. Strangley, this setup didn’t work.

We tried the same exact setup with a PC switchbox that also didn’t want to work with the HHKB. Our previous USB-hub solution worked with the switchbox also, but so did this Y-cable trick. Why it works with the switchbox and not the iPad is a mystery.

Back to DAC solutions, we read about Dr. Bott’s T3hub, which is a non-powered USB hub that supposedly “tells” the connected PC that it’s powered. This means that as long as there’s adequate power being pushed through a given USB port, the T3 “tricks” whatever mechanism is asking for extra power into thinking the power is there. We bought such a hub and connected it to the CCK, and connected the HHKB to the T3. We got an error that the device we connected needed more power, so we were again out-of-luck.

Just to see what would happen, we took our previous failed Y-cable setup but attached the T3 to the CCK, and plugged the one side of the Y-cable into the T3. With the power brick still plugged into an outlet, this setup worked, and actually makes for a more compact foorprint than the USB hub we used earlier, since we don’t need to carry around an extra power adapter. Unfortunately, this setup still doesn’t charge the iPad, and still requires an outlet.

Going portable.

We scrounged through our cable box one more time and dug out a portable battery charged via mini-USB. It was part of a kit to use an iPhone with a motorcycle, and basically provides additional power to the iPhone so you don’t deplete the iPhone internal battery as quickly when using GPS. Specs from the bottom of the battery are:

Capacity: 3200 mAh
In: 5V . 1A
Out: 5.5V

We charged the battery to capacity and then plugged our Y-cable into it. So on the one end of the Y-cable we had this battery and HHKB plugged in, and on the other end the Dr. Bott. The Dr. Bott was then plugged into the CCK, which was plugged into the iPad. The setup worked! This gives us a portable solution for the HHKB, and a simple USB extension cord gives us some room to move the HHKB further away from the iPad. The only downide to this setup other than the cable clutter is the lack of ability to charge the iPad with the CCK attached.

Our battery isn’t anything special, and we assume that any portable power source would work here. Just to verify, we again went through our cable box and dug out the requisite female USB-A adapter for our Solio rechargable battery.³ This worked just as well, with the added advantage of a solar recharging solution in exchange for a bigger footprint. Given the number of solar charging and battery solutions on the market nowadays, there are plenty of options here.

If you want to run the HHKB off two AA batteries, there are solutions available, or you could even build a MintyBoost if you want to go the DIY route. Our battery has five blue LED lights indicating battery charge, and with the keyboard hooked up in this configuration for a full work day, not a single light went out.⁴ Despite this, the actual power draw seems more than we expected, and may explain why Apple didn’t leave the allowable power draw through the CCK at 100 mAh.

Ultimately, our portable solution works, and keyboard battery life is about on par with the iPad itself. That’s not bad as long as you can deal with the hassle of charging a second device whenever you plug the iPad into a power source as well.

Alternatives?

The HHKB Pro 2 with a portable power source is the most compact Topre-based keyboard you’ll find to use with the iPad. The Realforce 87U is a larger alternative that may not require a power source, but the added keyboard size makes it less useful on-the-road. If you’re willing to forego Topre switches for mechanical switches like the Cherry MX series, a reasonable option is the KBT Poker, whose footprint is about on par with the HHKB Pro 2, and whose price is about 66% less. The downside is that the Poker’s build quality doesn’t meet that of the HHKB Pro 2, and the layout isn’t as ideal either, seeing as it doesn’t target the same keyboard-maestro audience the HHKB’s developers did.⁵ Fortunately, the Poker does come in various Cherry MX configurations, so you can get it in the very clicky blue configuration, the quieter tactile brown configuration, as well as the linear black and red configurations.[^5] This at least gives mechanical keyboard lovers several options, whereas the HHKB Pro 2 only comes in one Topre switch configuration with an actuation force of 45g.

For those willing to wait, you can occasionally find an HHKB Pro sold second-hand. These were produced prior to the HHKB Pro 2, and are basically the same keyboard, though lacking the integrated USB hub. Presumably, this means it would work with the iPad without any of the previously-mentioned accessory-hacking, though we can’t comment for sure since we haven’t tried one yet.⁶

For now, we’re quite happy with our workaround for getting the HHKB Pro 2 to work with our computing device of choice. The overall accessory breakdown, roughly, follows:

Accessory	Price
CCK	$30
Dr. Bott T3hub	$25
USB Y-cable	$4
USB extension cable	$3
3200mAh battery	$50

That comes down to roughly $107 on top of the cost of the HHKB Pro 2, so it’s not cheap if you don’t have some of this stuff already laying around. In our case, we only had to purchase the T3hub and Y-cable, so our costs weren’t that much as far as the equipment that ultimately made it into the setup goes.

Then again, if you love tactile keyboards, especially those utilizing Topre switches, you’re no stranger to paying more for quality. From our perspective, it’s worth it.